Last April, I explored listed airports and made the case for two Mexican airport groups:

• OMA, a domestic concessionaire of the Monterrey airport (46% of its revenue) and 12 other Mexican airports in the North Central region of Mexico.

• ASUR, the concessionaire of Cancún’s airport (57% of revenue), 8 other Mexican airports and a few other Latin American airports, including Puerto Rico and Colombia, and more recently added, Brazil, Costa Rica, Curaçao and Ecuador.

While the total return in dollar terms is positive since the investment 17 months ago, all three are well below the benchmarks, with ASUR performing the worst, returning just 6.6% versus the MSCI ACWI Index return of 41.7%.

Beyond the ongoing Middle East tension impact on global travel, there are a few other ongoing issues in the Mexican airport industry and the purpose of this article is to review how these issues might impact their earnings and free cash flow forecasts over the next five years, update our conclusions on their valuations and determine if we maintain (12-15% IRR), reduce (6-12% IRR) or exit (<6% IRR) the Mexican airports investment.

I had previously left out GAP, the largest Mexican airport group with 12 Mexican airports led by the Guadalajara Airport (25% of revenue) and two airports in Jamaica, from the initial Mexican airport investment case due to valuations, but with its recent share price decline, GAP now trades at valuations worth researching, and I include a detailed GAP valuation model in the article below.

GAP Valuation Model

Here’s a table detailing the key takeaways from the Mexican airports updated research.

Table of contents

1. The Mexican airport challenges: Analysis of key issues such as the Middle East tensions, MXN strength, ongoing Pratt & Whitney engine challenges and the potential Viva and Volaris merger.

2. Recent airport developments: Internationalisation of TAA, acquisitions at ASUR and GAP and the general state of Mexican airport passenger traffic.

3. GAP’s analysis and valuation: Passenger growth, ARPU, non-aeronautical revenue potential, integration of CBX and valuation model.

4. ASUR’s analysis and valuation: Passenger growth and ARPU, earnings impact of ASUR US Commercial and Motiva’s airport portfolio, debt levels, changes to the exit P/E multiple and valuation model.

5. OMA’s analysis and valuation: Monterrey’s passenger growth drivers in focus, ARPU across all airports, earnings potential and valuation model.

6. Portfolio allocation: Jenga IP’s exposure to Mexican airports and balancing contrasting risks and benefits.

1. The Mexican airport challenges

Why airports?

As a business model, airports generally meet what I look for in investable businesses. They are highly profitable (>30% EBIT margins), natural monopolies with high barriers to entry and benefit from global growth in tourism and business travel. Airports aren’t flawless, though. Their dependence on passenger volumes means they are still exposed to cyclical economic shocks, and there’s a limit to their value chain control due to regulatory and concession risks, and dependence on travel demand for locations served.

Why Mexican airports?

The best airports are those located in countries with growing GDP’s and populations, a strong tourism economy and a high diaspora population in nearby countries. Alongside Turkey and Thailand, Mexico is one of the few economies that fit this description, and the Mexican airports have the additional benefit of lower concession fees with good non-aeronautical opportunities through bonded warehouses (OMA), cross-border bridges (GAP) and duty-free and retail revenue growth (ASUR). As early movers in airport privatisation, they’ve also managed to acquire airport concessions in nearby countries, supporting their growth potential (ASUR and OMA).

With dividends included, an equally weighted portfolio of the three airports has compounded by 878% (USD terms) over the past 20 years, a 12.1% CAGR, versus the S&P 500’s return of 600% (10.2% CAGR) during the same period, while trading at one of their lowest valuations since going public today, 12x forward earnings.

It’s been far from smooth sailing for these three companies, and arguably, the current climate is one of the most challenging periods for them outside the financial crisis in 2008 and the global pandemic.

Before diving into each airport group, let’s first explore the overall industry challenges:

1. Middle East tensions: Airport revenue is directly tied to passenger volumes, so any factor impacting travel volumes impacts them. The Middle East tensions had an immediate impact on global travel as disruptions hit airlines around the world. The second-order impact was jet fuel prices, which nearly doubled by March, increasing the costs for air travel and leading to a few planes being grounded on routes that were unprofitable.

   

   In Mexico, Volaris, the second market leader, had cut flights by 9% in May due to higher jet fuel prices, while Viva and Aeroméxico, the other two airlines in Mexico’s oligopolistic airline industry, cut domestic flights by 1 and 3%, while prioritising more on international flights. Latin American travel has so far remained more resilient than other parts of the world, so this alone doesn’t explain the ongoing challenges.

2. Mexican peso strength: Airports earn international passenger charges in USD, and these international passengers also spend in USD at duty-free and restaurants, which means a stronger USD boosts their MXN earnings. The recent broader USD weakness, coupled with foreign speculation on Mexican interest rates and carry trades, has increased the MXN value against the USD, reversing losses from 2024.

   

   When these periods of heightened MXN strength occur, it also becomes more expensive for foreign tourists. Americans make up more than 60% of all tourists in Mexico, and given the higher Mexican prices, many have looked to alternatives such as the Dominican Republic for the summer holiday (more on this shortly). Among the three airport groups, ASUR, the parent of Cancún’s airport, is the most exposed to this and has seen passengers decline by 6% this year. GAP (Guadalajara) and OMA (Monterrey) have also experienced some weakness in non-aeronautical spending per departing international passenger, further highlighting the currency strength impact.

3. Failed World Cup expectations: Economists and sell-side analysts expected Mexico’s participation as the Men’s Football World Cup co-host to boost the Mexican economy, luring more tourists, travel activity and supporting the local economy. However, the impact from the World Cup was very limited to just Monterrey (OMA) and Mexico City. In fact, many tourist hotspots like Puerto Vallarta (-6% June passenger decline), Los Cabos (-1.9% June passenger decline) and Cancún (-11.5% June passenger decline) saw passenger declines, with some directly attributable to the World Cup, as potential tourists went to other destinations during the summer period.

“We had a romantic idea that we [Cancún] could become the hub of the World Cup. That’s how we worked, that’s how the actions and intentions were made.” - Rodrigo de la Peña, President of the Cancun Hotel Association.

Today, a few analysts have updated their models, now suggesting a sharp bounce back for the 2027 summer, and as I will discuss in our analysis for ASUR, I fear this is wrong.

4. Pratt & Whitney engine issues: Mexican airlines, particularly Volaris and Viva, which together represent 57% of the Mexican aircraft fleet, are among the worst-hit airlines from the Pratt & Whitney engine issues, present in many Airbus A319, A320, and A321neo aircraft. At the start of last quarter, both airlines had a combined 21%, 56 of 264 planes currently grounded, with Viva particularly impacting OMA due to the Monterrey flight routes and Volaris more exposed to GAP’s airports like Tijuana and Puerto Vallarta.

5. The Viva and Volaris merger: The proposed merger between both airlines will have positive and negative effects for the airports. For flight routes they currently both serve, we will likely see route optimisation in the long term, with less frequency when compared to their current state, leading to lower landing and parking charges for the airports. For passenger growth, though, the most important revenue variable for airports, management argues, and I suspect the larger economies of scale gained from parts sourcing could lead to slightly better prices for customers, thus increasing access for Mexicans, which in turn boosts passenger growth and fees for airports.

Of these five, the Middle East tensions and Pratt & Whitney engines are temporary supply shocks which should lessen in impact over time, while I also suspect the World Cup disruption to tourist locations and MXN strength are temporary demand-side shocks for the airports. The true long-term challenge here is the Viva and Volaris merger, and the fact that this issue could turn out to be a positive factor for the airports keeps me optimistic there could be some light at the end of the tunnel for the airports.

Next, let’s look into each of the three airport groups, discussing the growth drivers, individual challenges and how I currently view their valuation and potential shareholder return.

2. Recent developments

A few factors have changed since my initial deep dive into Mexican airports that require a fresh perspective on their prospects and valuations.

I. Technical assistance agreement (TAA): At the time of their privatisation in 1998, Mexican airports needed technical partners to support operations and Capex plans. They created a separate share class for these partners and also paid them a fee, between 2.5% and 5% of operating profits per year. Over time, current controlling shareholders like Fernando Chico Pardo of ASUR and Laura Díez-Barroso Azcárraga of GAP acquired shares in the TAA partners, ITA and AMP respectively, meaning that minority shareholders were also paying the majority shareholder for technical assistance.

For the initial years, this arrangement made sense, but today, these airports are international players themselves, have the expertise and no longer need support to plan and execute Capex programs, and to terminate this arrangement, both ASUR (2.4% dilution) and GAP (3.2% dilution) issued shares to acquire the TAA and, in return, will not pay any more TAA fees going forward. I have reflected these in the updated valuation model, as this reduces the cost base and increases profit margins for both airport groups.

OMA hasn’t yet announced a similar deal and still pays the TAA fee to Vinci, but I suspect they will follow suit in the coming quarters.

II. Acquisitions: Although airports are generally the least acquisitive of the various transport infrastructure groups, 2026 has been the most acquisitive since inception for Mexican airports, with GAP acquiring the remaining 25% it didn’t own of CBX (Cross Border Xpress), the cross-border bridge that links the Tijuana airport (12% of GAP revenue) with the U.S. border, while ASUR closed the acquisitions of URW Airports, now called ASUR Airports, an American concessionaire of non-aeronautical and retail operations at terminals in Los Angeles aiport (LAX), Chicago O’Hare, and JFK airports, and Motiva’s airport assets, a concessionaire of 17 airports in Brazil, and 1 in Ecuador, Costa Rica and Curaçao.

There’s a lot of detail to explore in these acquisitions, but for investors, the most important question is how attractive they are and what the potential net earnings impact these acquisitions will contribute to the companies.

GAP’s acquisition of CBX

GAP’s CBX asset is the most attractive of the three acquisitions for three main reasons. First, CBX is a highly profitable asset with EBITDA margins estimated at 65% and strong conversion to free cash flow. Second, while CBX has alternatives for travellers between the Mexican border and the U.S., it’s the lowest-cost travel solution.

American travellers to Los Cabos, Mexico, from California would find it cheaper to travel from CBX into Tijuana (US$165 round-trip price) than to fly directly from LAX (US$342, 3-hour drive from CBX) or San Diego airport ($342, 30-minute drive from CBX). As a result, since 2015, CBX has steadily taken share from other Californian airports, most particularly San Diego; 58% of its 4 million passengers came directly from San Diego, and I believe this has further room for growth.

Finally, CBX is a fully USD asset, and while MXN exposure is more lucrative for investors today, it will come in handy during periods of MXN depreciation.

Overall, I believe CBX can contribute Ps. 3.05 billion in additional revenue for GAP, a very conservative 3.5% revenue CAGR between 2027 and 2030, while maintaining its 65% EBITDA margins, and I factor this into the valuation model. Also important for GAP is how this shapes the revenue potential for its Tijuana airport.

ASUR’s Motiva airports and URW acquisitions

Although sell-side analysts are yet to reflect the addition of both acquisitions in their models, I have reflected both in our ASUR model, and there are some key points worth discussing.

Motiva’s acquisition

For some background, Motiva, previously called CCR Group, is the largest listed transport infrastructure group in Brazil and had diversified away from toll roads into airports (45 million passengers) between 2012 and 2022, starting with Ecuador’s Quiport (5 million passengers). However, their acquisitive growth came with high debt, as many of these airport concessions came with high initial payments.

Motiva’s net debt levels had reached a high of R$34 billion on a market cap of R$25 billion, while spending a third of its EBIT on interest payments. Toll roads are more profitable than airports, and after a review, Motiva concluded it needed to refocus on toll roads and ground transportation and initiated the sale process of its airport assets.

This background is important because ASUR isn’t acquiring a Greenfield airport or bidding for a new concession. Here, it’s buying an existing concession operation, and while attractive from a revenue perspective, it comes with a high price for ASUR’s debt levels going forward.

My emphasis on the debt incurred from the deal is due to where ASUR is coming from. Historically, ASUR ran the most conservative airport balance sheet in the world; it maintained net cash positions and paid a big dividend in 2025. However, that same balance sheet will jump from net cash in 2024 to Ps. 58.4 billion (45% of ASUR’s market cap) at the end of 2026 (see chart above). More critical for ASUR’s operating earnings is the interest the new debt comes at.

Some of the new debt, such as the infrastructure debentures with the Central Bank of Brazil, are priced at inflation +6.96 - 8.05%, incurring much higher interest costs than ASUR’s existing debt.

At face value, the 9.4x EV/EBITDA seems like a good deal, but as I dug deeper into Motiva’s balance sheet and accounts, I concluded Motiva may have had the better side of this deal and could have long-term consequences for ASUR shareholders.

URW U.S. terminals acquisition

In December 2025, ASUR closed the acquisition of URW’s commercial concessions at 9 U.S. airport terminals, 6 in LAX, 2 in JFK and 1 in Chicago’s ORD airport. These concessions are purely non-aeronautical revenue; ASUR earns a royalty from retail spending at these terminals and then pays the airport a tenant rent. These fees are unregulated and tend to grow faster than aeronautical revenue but come with higher costs. Rent fees alone are 85% of the previous year’s revenue, minus some expenses at some terminals.

ASUR paid Ps. 5.63 billion for URW’s U.S. airport terminal assets, and while the overall deal price looks particularly high given the low EBITDA expectations for 2026, potential growth in LAX and the opening of the JFK New Terminal 1 in 2027 could ramp up revenue and EBITDA for the next few years, and I suspect the deal is closer to 10-11x 2028 expected EBITDA. Not cheap, but it comes with less debt when proportionally compared to the Motiva acquisitions. It’s also important to note that its revenue stream is 100% USD, which we view as long-term favourable for ASUR.

From management’s perspective, I do understand the rationale behind both deals. Colombia’s airport concession will be coming to an end in a few years, with the regulated revenue portion ending in 2028, and Cancún, as we will discuss, may have reached its growth peak, an overall situation that neither OMA nor GAP faces. As a result, there’s a lot of pressure to diversify, reduce exposure to tourist passenger flow and increase their bargaining position for future concessions. These goals need to be balanced with great deals, as they did with Puerto Rico and Colombia airport concession acquisitions.

III. Mexico’s airport passenger flow

Before diving into the valuation models and assumptions behind revenue, costs and passenger growth, it’s always useful to put the current climate into perspective. Prior to 2019, the airports shared a similar growth model:

• 5% passenger growth

• 6% pricing growth

• An additional 1- 2% from operating leverage

These led to a 13% EPS CAGR, but in recent years, revenue per passenger grew slightly faster to 7-9% due to the MXN depreciation, as international passengers became a larger share of revenue and tariff pricing renewals, while passenger growth declined from 5% to 3-5% as the discount airlines, Volaris and Viva, gained more scale, leading to slower incremental growth share.

The table below highlights this dynamic across passenger volume, revenue per passenger and total Mexican revenue. I also included the passenger growth for 2026 (Jan - Aug), which helps put the points discussed earlier into perspective.

As you see from the table, ASUR grew its passenger volume the slowest (Cancún weakness in 2025), but its overall revenue grew the fastest between 2019 and 2025 (USD appreciation impact). For 2026, the trend continued, again driven by weakness in Cancún.

Year to date, OMA has so far been the best performing from a passenger growth lens, particularly driven by Monterrey (46% of OMA’s revenue) and its smaller airports like Chihuahua, Durango and Tampico.

I expect the weakness across the three airport groups to continue for the rest of 2026 and possibly the first half of 2027, with the rebound mainly driven by when the Middle East tensions are resolved and normalisation of the Pratt & Whitney engine situation for Volaris, followed by Viva.

The next table highlights the current passenger growth for the main airports within each airport group:

• GAP (Guadalajara, Los Cabos, Tijuana and Puerto Vallarta)

• OMA (Monterrey, Culiacán, Ciudad Juárez, Mazatlán and Chihuahua)

• ASUR (Cancún, Mérida and Villahermosa)

As shown in the table, GAP is much more diversified than the other two groups, with half of its revenue coming from 3 airports, rather than just one airport as with ASUR (Cancún) and OMA (Monterrey). GAP’s big three airports are also diversified by purpose (business, family and tourism travel) and is the core reason why GAP deserves a higher earnings multiple than its peers.

An important point for our earnings growth model, though, is the current state of passenger growth across the airports, and as highlighted in the final column, there are two interesting trends worth discussing.

• Tourism weakness (Cancún, Los Cabos and Puerto Vallarta): Generally, I find tourism to be the highest growth-potential purpose but also the most exposed to consumer trends and the economy. During boom times, as we saw in 2022 and 2023, these perform well, but suffer during consumer weakness and also face a bigger substitution risk. GAP’s Puerto Vallarta, a popular destination for American retirees, was particularly hit by security concerns from cartel activity earlier in the year and Spirit Airlines’ bankruptcy and resulting capacity loss.

  

  I don’t expect an immediate recovery here, given these are mid-term issues and will take some time to recover. Cancún, on the other hand, suffers from a more long-term issue. For years, Cancún served as the undisputed number 1 tourist destination for Americans outside the U.S., and while it’s still leading the market, I think analysts could be underestimating the threat other tourism destinations pose, particularly given the MXN strength. Take Punta Cana in the Dominican Republic (DR), an alternative destination for American tourists. Between 2018 and 2025, Punta Cana’s passenger flow as a share of Cancún’s grew from 31% to 37%, and for the first 6 months of 2026, its share is at an all-time high of 43% (see graph below).

Beyond Punta Cana, other destinations like Aruba, Bahamas, and Curaçao each reported growth for H1 2026, albeit with some weakness for American tourists. Without MXN depreciation, Cancún will remain at a price disadvantage, and further supporting this view is our most recent study of price differentials between Cancún and other alternatives.

Cancún pricing study

Given that the majority of Cancún’s American tourists come from Miami, Dallas and Houston, with the support of AI, I web-scraped current pricing for an all-expense, 7-night trip between the three destinations to Cancún and compared it to 21 alternatives across Mexico, the Caribbean, Turkey, London and Miami (without flights) on peak dates in June and December 2027.

As you can see from the table above, Cancún, as well as other Mexican alternatives, has increased in total pricing (hotel, flight and day spend) and once seen as one of the lowest-cost beach travel destinations, is now priced (USD 1,925) above alternatives like Punta Cana (USD 1,859, USD 66 difference). A holiday package in Miami, without a flight ticket, is increasingly becoming nearly as cost-effective for Americans when compared to a Cancún holiday. Without MXN depreciation, Cancún’s appeal is unlikely to bounce back to what it was, and I factor this into the ASUR valuation model.

A final blow and more unpredictable factor for Cancún has been the presence of Sargassum seaweed, a smelly microalgae across the Riviera Maya beaches. 2026 is set to be the worst year on record and, according to government data, the region picked up more Sargassum in the first half of 2026 than the whole of 2025 (see chart below), and its growth on Cancún beaches leads to hotel closures, lower occupancy rates and less tourism demand. We have no sense of how this factor changes for the future, and for valuation and margin of safety, I assume this stays in a mild-to-bad state.

• Business travel strength (Monterrey, Guadalajara and Villahermosa): Business travel has been particularly strong outside Mexico City, with Monterrey being the biggest beneficiary as corporates set up manufacturing hubs there to take advantage of its proximity to the U.S. FDI increased 22.4% for the first half of 2026 in Monterrey and, for the first time, represented more than 10% of Mexico’s total FDI, and I continue to see strong evidence of manufacturing activity supporting business travel. That said, it’s always important to note not to become overindexed to one trend, no matter how strong it looks, and this is why I still view Guadalajara as the higher-quality earnings given the diversification across business, tourism and passengers visiting friends and family. As a result, this also keeps Guadalajara more exposed to international passengers (6 million versus 2.5 million PAX).

Villahermosa (and Veracruz, ASUR), on the other hand, is a more unique business travel boom, highly correlated with oil and gas activity in Mexico and often experiences travel boom and bust cycles. Lately, there’s been policy by the Tabasco government to support domestic flights for the petroleum sector, and I believe the 9.2% growth experienced in 2026 YTD is directly attributable to this. Given the cyclical nature of oil and gas activity, I wouldn’t expect recent growth trends to simply grow forever, and I factor these cycle turns into our ASUR valuation model.

This leads us to the valuation model and potential IRR for GAP, OMA and ASUR.

3. GAP’s valuation model

GAP Valuation Model

You can view the GAP valuation model here.

Rather than discussing each line in the valuation model, I will provide an overview for each company, discussing key passenger growth & non-aeronautical revenue, costs, and earnings considerations for each Mexican airport group. Also note that explanations for key growth lines are directly included in each valuation model spreadsheet.

GAP passenger growth and non-aeronautical revenue

• Passenger growth: Factors such as Guadalajara’s ongoing business travel boom, recovery in Los Cabos and Puerto Vallarta (assuming improved security concerns), mild growth in Tijuana and recovery in Jamaica’s affected tourism flow from Hurricane Melissa, GAP’s total passengers could grow by 3.95% between 2025 (63.7 million) and 2030 (74.9 million), with Guadalajara contributing to 49.3% of the total 11.2 million passenger growth.

• Revenue per passenger: I expect a reduced pricing pace for GAP’s revenue per passenger to 4.7% annually, led by strong pricing potential in Guadalajara and the tourist airports (Los Cabos and Puerto Vallarta) and Tijuana.

• Non-aeronautical revenue: Excluding the CBX acquisition, non-aeronautical revenue could grow by 8.2%, supported by strong potential across cargo warehouses, food and beverage and car rental categories. I estimate that the CBX acquisition adds an additional MXN 3 billion in annual revenue (17.5% of non-aero revenue) by 2030.

• Total revenue growth: Altogether, this should lead to total revenue growth for GAP of 9.7% between 2025 (MXN 32.5 billion) and 2030 (MXN 51.7 billion).

GAP costs and earnings

• I expect costs per workload to increase by 5.3%, versus revenue per workload growth of 6.2%, leading to some operating earnings margin expansion (54% to 56% EBIT margins) driven mainly positively by the removal of the technical assistance fee, which historically averaged 7% of GAP’s total costs and the high-margin CBX acquisition.

• Wage inflation could be higher given the 15% increase in minimum wages in Jamaica, security staff for the bonded warehouse and cargo businesses and others.

• Interest expenses: I expect material acceleration in GAP’s debt level, peaking at 2.5x net debt/EBITDA in 2027 with a potential 8.3% average financing cost. Overall, this slightly compresses net income margins from 29.4% in 2025 to 29.3% in 2030 and an overall 9.67% net profit CAGR.

GAP’s Valuation

I believe 19x P/E is the appropriate multiple for GAP (see image above), driven by its diversified airport and traveller profile, currency balance between MXN and USD streams, and attractive margin and capital return profile.

Combined with MXN 42 billion in dividend payments, GAP IRR could amount to 11.7% between now and the end of 2030 (4.33 years).

4. ASUR valuation model

Read more

Intuit reported its FY 2026 earnings results yesterday, and despite the negative market reaction, I was generally pleased with both its FY 2026 results and guidance for FY 2027.

I will discuss my thoughts and analysis in this article, but before I dissect the results, I want to first mention three management decisions Intuit made that were quite good.

1. Mailchimp: Mailchimp was a costly acquisition and management historically buried its performance in the annual investor presentation, rather than the more important SEC filings. The recent decision to be more explicit on Mailchimp’s results shows Sasan Goodarzi, Intuit’s CEO, and the team are taking a more accountable stance with past capital allocation decisions. This culture will certainly reward shareholders over time.

2. Stock-based compensation treatment: One of my biggest frustrations with U.S. software companies was their exclusion of stock-based compensation from Non-GAAP earnings calculations. I addressed this issue during the Intuit deep dive (see page 74 of 80) and explained why I account for this in my own non-GAAP calculations.

Intuit’s management confirmed it will now reverse its decision to exclude stock-based compensation, aligning better with all shareholders.

3. Growth ambitions: During my months-long research into Intuit, I was incredibly perplexed about why management thought 20% revenue growth was possible. I reckon they simply needed a justification for Intuit’s previously extremely rich valuations two years ago. Here are its CEO’s own words during the FY 2025 Investor Day.

“We want to be one of the most trusted companies in the world, and we aspire to 20% top line growth….The second and last thing I would touch on this page is Intuit grew 16% this past year. And internally, we believe 20% is possible. It’s our aspiration.” Sasan Goodarzi, FY 2025 Intuit Investor Day.

Investors who looked deeply into each of Intuit’s businesses would quickly conclude that 20% would be near-impossible to achieve, so its good management has now backtracked from this target to a more realistic growth range. Intuit doesn’t need to grow 20% to achieve a 20% IRR.

A recap of Intuit’s investment case

Earlier in June, I made Intuit an 8% position in the Jenga IP Global Equities portfolio and the investment case was predicated on:

• Revenue: 11.9% 5-year CAGR

• EBIT: 12.4% 5-year CAGR, with 2030 EBIT margins of 26.8%.

• Net income: 12.7% 5-year CAGR

• Shareholder returns: $25.9 billion net shareholder return between FY 2025 and 2030 (dividends and repurchases net of stock-based compensation).

• Earnings multiple expansion: An exit multiple of 23x by FY 2030 supported by Intuit’s business quality and growth prospects.

Altogether, these points to a potential IRR of 20.3% from our purchase price of $84.9 billion market cap or $269.1 per share.

The important question I ask myself every quarter and fiscal year is: Is Intuit on track to achieve our investment target?

Let’s first dig into its 2026 results to answer this question.

1. 2026 financial estimates versus actual

As shown in the table above, Intuit performed better than my estimates in every key financial metric, and while my revenue estimates were close, I overestimated every actual cost line performance:

2. 2026 cost estimates versus actual

Even with the $293 million restructuring charge from the ongoing layoff of 17% of the workforce, Intuit still beat its own GAAP operating income guidance of $5.82 billion, and it’s impressive that we see such cost management even as it chases expensive AI ambitions across QuickBooks and the lower-margin TurboTax Live business.

I had included a margin of safety across all cost lines, but it’s evident that my cost pressure worries are overblown, especially when we look to the 2027 operating income guidance. More on this shortly, but let’s next dig even deeper into key Intuit KPIs to gain a better understanding of FY 2026.

3. Division revenue estimates versus actual performance

It’s important to look under the hood when assessing revenue growth, and for Intuit, there are three segments that particularly matter for the investment case: Online Accounting (QuickBooks), Online Services (Bill Pay, QuickBooks Payments & Capital, Payroll, and Mailchimp) and TurboTax.

As you see from the table, QuickBooks Accounting was ahead of my estimates, and while we don’t have a complete perspective (more information will be provided during the FY 2026 Investor Presentation), a larger share of the 22.6% revenue growth was driven by pricing increases. As we discussed during the deep dive, Intuit raised prices by 15-17% in 2025 for its core QuickBooks plans (see table below).

Paying customers grew by only 3% to 8.9 million in FY 2026, below my own estimate of 9.2 million paying customers for FY 2026 (see page 67 of 80 in the Intuit deep dive).

In the future, price increases will play a smaller role in QuickBooks growth, and as management discussed, it will focus on using QuickBooks Free and Lite to grow its user base, with the hope that they convert into higher ARPU customers once they add Bill Pay and Payroll as their businesses scale during the earnings call.

“A key component of our new-to-the-franchise strategy is widening the front door with QuickBooks Free and QuickBooks Lite. These offerings create low-friction entry points to reach millions of businesses earlier in their journey and build a relationship with them from the start.” - Sasan Goodarzi, FQ4 2026, Intuit Earnings call.

The other core growth driver here is Intuit Enterprise Suite, its newer mid-market ERP and accounting solution for larger customers. Intuit reported the mid-market grew by an impressive 28%, with 3/4 of new customers coming from existing customers who either upgraded from QuickBooks Desktop or Advanced plans.

A large share of these customers are businesses in the American construction sector, and Intuit still have a lot to do to prove it can actually grow its mid-market solutions beyond them.

Online Services division growth came in slightly below my estimate. From the earnings call, we know QuickBooks Payments exceeded our own 26.5% volume growth to $220 billion:

“We are seeing strong evidence of this today with significant opportunity ahead. Businesses manage over $2.7 trillion in invoices through QuickBooks every year, and total online payment volume, including bill pay, grew 30% to more than $225 billion for the full year. And as customers grow with us, they adopt more of the platform.” Sasan Goodarzi, FQ4 2026, Intuit Earnings call.

However, little was said about Payroll, and from the overall picture, Payroll was the disappointment here, relative to my estimates. I will watch out for further commentary during the investor presentation next month.

Let’s next move to the main problem within Intuit, TurboTax.

TurboTax DIY issues

There’s no escaping the fact that TurboTax DIY is Intuit’s big problem amid increasingly more severe competition. Management hoped tying Credit Karma with TurboTax could curb customer losses and churn (Credit Karma customers represent a large share of the lower-income households), but its customers are increasingly becoming less loyal to tax filing software and shopping more and longer for the lowest-priced tax solution ahead of the tax season.

This means Intuit will likely have to spend even more on marketing, customer discounts and Credit Karma cross-selling to curb these TurboTax DIY losses.

Overall, 2026 was quite good for Intuit, and the next question is how does management view performance for 2027?

4. 2027 estimates versus management guidance

At first glance, the 2027 guidance might seem like a cause for concern, but once we factor in the addition of stock-based compensation to non-GAAP calculations, Intuit’s 2027E net earnings per share growth of 23% is very attractive and even exceeds my 2028E diluted EPS estimate of $20.09, while 2027E GAAP operating income guidance similarly exceeds my 2027E and 2028E, nearly approaching my 2029E modelled EBIT of $7,766.7 million.

When I compare management’s mid-point guidance for FY 2027 with my own 2027 estimates, the challenges lay in Global Business Solutions and TurboTax’s slower than anticipated revenue growth.

While I’m generally more focused on free cash flow and earnings growth, revenue shortcomings are also important for any long-term investment case. The difference between my total revenue estimate and management’s is $331 million, a 1.4% difference, and while I don’t think this is a big enough gap for a structural worry, I’m aware revenue shortcomings should certainly be paid attention to.

For coming quarters, I will be watching QuickBooks paying user growth (can it return to a >5% grower?) and TurboTax DIY (can overall TurboTax return to a 4% growth rate?), particularly in FQ3 2027.

Conclusion and final thoughts

The key worry for Intuit is whether AI replaces both QuickBooks and TurboTax.

For QuickBooks, Intuit’s results show users still see value in its platform. Not many legacy software companies can grow users by 3% after raising prices by 17% in a year. Secondly, payments’ 30% volume growth proves Intuit’s strategy of “beyond just bookkeeping” is taking shape and will materially drive revenue per user over time.

For TurboTax, I remain watchful on its DIY defensiveness to both cheaper alternatives and AI empowering the competition more than Intuit.

For now, the overall picture still looks very healthy despite the market reaction, and I remain long Intuit and still believe it could match our 20% IRR target over the next 5 years.

Paid readers can access our initial Intuit valuation model below the paywall.

I published a deep dive into Adobe last week and you can read the abridged one-page summary below.

Read more

An easier to read PDF copy is available just below for paid readers.

Adobe, the world’s leading creative software company, has been a tale of two halves in the past decade. The transition from perpetual licence to subscription SaaS promised a much brighter future, business model and resilience. Investors stayed ever bullish, pushing its price-to-earnings to a record 70x P/E in late 2020.

However, the same technological force that once promised more growth for Adobe now threatens its value proposition to creatives, as the most advanced software for content manipulation. The same company, once priced at 70x P/E, is now valued at 11x P/E.

I myself have also gone through two phases in my research into Adobe. I first started extremely sceptical but also eager to learn about how Adobe had lasted so long, just as we did with some other foundational five software companies like Microsoft and Intuit.

However, as I researched Adobe more deeply, poring through decades of company accounts and business history, speaking with many customers and doing general research, I gained three important insights about Adobe. First, it’s one of the few software companies with “horizontal industry standards” status. Second, there’s a key difference between “content manipulation” and “content generation”, although they meet in certain applications. Third, the power of product bundling is extremely important in valuing software companies.

• Market capitalisation (As of 17th June 2026): $82.4 billion

• Jenga IP 2030 FY estimated market cap: $194.2 billion

• Potential IRR (including buybacks): 25.3% IRR

• Jenga IP Quality Rating: 75.3/100

I will discuss these three points throughout the Adobe deep dive, but before getting into the analysis, I also want to highlight 5 key points that I believe are misunderstood about Adobe.

1. AI will replace Adobe: Not true. As we’ll discuss, AI and Adobe aren’t mutually exclusive and Adobe has multiple ways of monetising AI capabilities while also leaning into its trusted and secure creativity and marketing software.

2. Ongoing customer revolt will ruin Adobe: This isn’t the first nor second wave of customer revolt against Adobe. The same customers once outright hated and petitioned against its subscription and cloud transition. Or we can point to when Steve Jobs publicly slammed Adobe Flash. Adobe has dealt with this several times and survived.

3. All subscriptions are equal: Far from reality. The biggest challenge Adobe faces compared to peers like SAP and Microsoft is its subscription resilience when customers face recessions. Marketing and advertising budgets are the first to go, directly impacting Adobe’s seat count. This is the main reason Adobe should never have been valued at 70x P/E.

4. Figma will replace Adobe: Given that Adobe offered a quarter of its current market cap ($20 billion) to buy the creative product design startup, one might conclude Figma will soon replace Adobe. I don’t believe this is true, and as I’ll discuss, there’s a world where Figma’s collaborative app and Adobe’s fleet of design solutions co-exist.

5. Adobe is a one-product company: Far from reality, too. Possibly no other software company has reinvented itself as much as Adobe in the last 40 years, transitioning from the original PostScript software for publishers to over 100 current products.

Table of contents

1. Adobe’s history: A look into key innovations (PostScript, Acrobat, Photoshop and Illustrator) that shape Adobe’s moat today, case studies on economic recessions and the cloud subscription transition, acquisitions and the current state of its barriers to entry.

2. Adobe’s product suite: Insights into the Creative Cloud products and plans, document tools for Acrobat and the marketing software solutions. Insights into their competitive landscape, contribution to the Adobe ecosystem and customer persona analysis.

3. The creative and marketing software industry: Analysis of the raster image and vector editing markets, broader desktop publishing and the customer experience market. Insights into generative AI tools and the all-in-one product design market.

4. Adobe’s AI challenge and opportunity: An assessment of the current state of Adobe’s AI threats, insights into Adobe Firefly’s business model and value chain, customer use cases and growth of AI-generated media content and a comparison of Adobe’s competitiveness relative to Google and OpenAI.

5. Adobe’s business economics: An overview of Adobe’s revenue by segments (digital media, digital experience and publishing and advertising), revenue by customer group (creative & marketing professionals and business professionals & consumers), revenue by region and an analysis of its cost structure and profitability.

6. The growth potential: A discussion on Creative Cloud’s user growth, pricing mix and Firefly monetisation, Acrobat’s user growth, marketing solutions customer cross-selling opportunity and potential acquisitions.

7. Adobe’s risks and challenges: Further discussion on Adobe’s AI challenge, state of its pricing power, management challenges, exposure to economic challenges, churn and the competitive climate.

8. Adobe’s valuation: Adobe’s net earnings and free cash flow projection to FY2030, valuation analysis and IRR projections over the next 5 years.

9. Conclusion: Final thoughts on the AI disruption to software companies and my portfolio management approach with software companies.

1. Adobe’s History

“Technology is never created in a vacuum. If you’re working on it, then someone else is too. The only way to succeed is to get there first.” - Charles Geschke, Adobe’s co-founder.

Every investment analysis I make starts with two questions. First, if given $1 billion in cash with a team of the brightest Silicon Valley technology talents, how easily can you replicate the product, value proposition and win their customers? The second question is, if the world’s most innovative companies (Google, Apple and Microsoft) were interested in your industry, how easily can they erode your market share?

Luckily and unluckily for Adobe, they’ve had to face the reality of both questions over the last 40 years. Figma, an indirect competitor in the collaborative product design industry, raised over $1 billion. At the same time, Apple, Microsoft and Google each compete with different Adobe products, and in many cases, offer customers a free version as an alternative.

For the second time since going public in the early 1980s, there’s a deep worry among market participants about the outlook for Adobe, its survival and future prospects. To make the case for Adobe’s resilience in today’s AI world, we must first study its history through the lens of four products that shaped Adobe as it is known today. These are PostScript, Portable Document Format (PDF), Photoshop and Illustrator.

PostScript (1983)

A technology challenge in the wake of the PC revolution in the early 1980s was printing. Printers and computers spoke different languages, which meant formats and fonts changed depending on the printer used. Adobe’s co-founders, John Warnock (later its CEO) and Charles Geschke (its COO), initially developed a solution while working at Xerox’s PARC, but difficulties with getting internal approval for their commercialisation forced them to start a new company, Adobe and develop the PostScript software. Rather than simply reprinting in pixels, PostScript used mathematics to instruct printers, ensuring shapes, fonts and formats were maintained regardless of the printer used.

The nature of a ‘standards business model’ where one product sets the approach for the whole industry means there’s a real first-mover advantage, and after considering different business models, an investment from Apple and Steve Jobs, alongside a license agreement for 5% of sales from Apple’s LaserWriter printers, Adobe settled on a royalty business model which turbocharged its revenue from day one.

Adobe was already profitable in its first full operational year in 1984, with revenue growing from $2.2 million to $16 million in two years (see image above). Competitors like Xerox’s Interpress (created by Adobe’s co-founders while at Xerox) and Imagen, then utilised by HP, meant Adobe knew it needed to grow fast and decided to go global, starting with an agreement with Morisawa in Japan for its fonts business, where customers similarly paid for licenses to install Adobe’s fonts.

Had Adobe stopped innovating here, it would have become another failed startup. Today, PostScript itself is immaterial to Adobe, sitting within the publishing and advertising division, representing less than 1% of Adobe’s revenue. However, it led to a pivotal transformation: the creation of the Portable Document Format (PDF) and Illustrator.

Illustrator (1985)

John Warnock, one of Adobe’s co-founders, noticed that his wife, Marva, a graphics artist who later designed Adobe’s initial logo, had work that paralleled PostScript’s technology and thought of extending the technology to the design world. To achieve this, he recruited PostScript engineer Russell Preston Brown and some in-house designers to develop the Illustrator software, launching in March 1987 for $495.

Scepticism from the design world was initially high, and Illustrator faced competition from CorelDRAW and FreeHand by Aldus (Adobe would later acquire Aldus in 1994), but over time, especially after Microsoft and Apple launched their partnership, ‘TrueType’, commencing the font wars in the process, Adobe’s diversification efforts were rewarded.

During the Seybold Seminars trade show, when Bill Gates announced the TrueType partnership to curb Adobe’s font monopoly, John Warnock responded at the event:

“The biggest bunch of garage and mumbo jumbo…We are going to out-invent the bast***d”

Just as Adobe continues to diversify beyond the Creative Cloud division today, its priority in the early 1990s was diversifying from PostScript and its font royalties and then a recent recruit, Fred Mitchell, identified the software application market as the main strategy.

Photoshop (1990)

“I ran, not walked, to Mr Warnock’s office and said, ‘We should buy this [Photoshop].” Inside the Publishing Revolution

Unlike PostScript and Illustrator, Photoshop wasn’t created internally at Adobe. Fred Mitchell came across the Knoll brothers, who developed Photoshop during Thomas Knoll’s PhD thesis. Unlike Illustrator, which was built on vector-based editing and enabled artists to draw line art from scratch, Photoshop enabled photographers to work on images from other sources.

Although optimistic, Adobe, like many others, underestimated the success of Photoshop and had bought Photoshop via royalties on units shipped. As time proved, significantly more people capture photographs than draw vectors, and in less than two years, Photoshop overtook Illustrator in sales, and then launched on the Microsoft Windows OS.

Over the next few years, Adobe acquired and established other products. One among these that remained influential for decades was Acrobat, an application for creating and viewing documents in the PDF file format.

Portable Document Format (1992)

Building on PostScript’s technology but with varying technical differences, Adobe diversified into solutions for the paperless office. Codenamed “Project Camelot”, the PDF brought an alternative text file format, allowing users to engage with an uneditable, mailable and printable document. Unlike PostScript, though, the PDF itself wasn’t initially monetised, and a year later in 1993, Adobe created Acrobat, a software application targeted to businesses for viewing these PDFs.

Unlike the other applications, Acrobat and the PDF were slow to gain traction, and it wasn’t until the internet that its growth accelerated. Early adopters like the U.S tax IRS for tax returns, advertising agencies, the Associated Press Online Service and then later financial institutions with the Acrobat 4.0 release in 1999 spurred the initial rise of PDFs.

Today, these four products gave rise to Adobe’s digital media division (76% of revenue) and to complement them, various acquisitions like Aldus Corporation in a stock swap deal then valued at $500 million in 1994 consolidated Adobe’s position in desktop publishing. Today, Adobe’s InDesign and After Effects, two niche creative products, still contribute to the Creative Cloud offerings.

Acquisitions remained a key part of Adobe’s business strategy, and rather than simply empire-building, most acquisitions played important roles in shaping Adobe’s product scope today. The table below highlights 12 major Adobe acquisitions and their surviving products. As you can see, even products acquired 30+ years ago remain integral to Adobe’s ecosystem.

Although the history recap above simplifies the Adobe story during its initial years, the period was broadly positive for Adobe, despite the challenges from Apple, Microsoft and many peers in desktop publishing. Adobe’s revenue and operating profits grew every year until 1996, when there was its first profit slowdown, marking the first competitive threat: the internet.

Test of time case study 1: The internet (1995 - 2002)

“At the peak of the internet bubble we were on the outside looking in.” Bryan Lamkin on Adobe

Every new technological shift poses some threat to existing technology leaders, and similar to AI, investors questioned if Adobe could survive the internet age, with its shares falling by 61% between December 1995 and September 1998. At that time, Adobe was already a major software company, only behind Microsoft in market share and was the 6th most profitable publicly listed software company.

Compared to Microsoft, though, Adobe was less prepared for the internet, which attacked its business on all fronts. Two, however, stood out: product relevance and employee morale.

The internet posed a direct threat to Adobe’s PDF as HyperText Markup Language (HTML) increasingly became the standard for web publishing. As HTML grew, Acrobat struggled, and despite attempts to enter the HTML design market with PageMill, it wasn’t until its partnership with Netscape’s web browser that PDF users could click web links directly on PDF documents that Adobe curbed PDF’s relevance loss.

Arguably the more critical challenge was the lack of employee morale. As Adobe’s shares fell while startups sprang up with attractive stock options for employees, Adobe employees became frustrated with internal growth, leading to lots of infighting. To deal with this issue, Adobe’s co-founders concluded that turning Adobe into a lean organisation was the best solution. They created a new team led by its eventual CEO from 2000 to 2007, Bruce Chizen, who in turn assembled a new team led by Shantanu Narayen, Adobe’s current CEO.

With its more focused organisation and clearer strategy, Adobe doubled down on its approach to new technological shifts, which was centred on (1) adapting existing products to new technologies, (2) using acquisitions to support adoption and stickiness. Acrobat was a good case study of this. After years of being offered as a free version, the internet unlocked a monetisation path for Adobe, and by 2001, Acrobat became the standard tool for enterprises, representing 24% of Adobe’s revenue. As of 2000, two challenges Adobe hadn’t yet faced were management change and surviving a deep recession.

Test of time case study 2: Management change

Another key sub-category within test of time is surviving the founder transition. Until 2000, Adobe was led by its two co-founders, and when founders step back, it’s often common for key employees to similarly resign, morale to decline and for the company to lose its product and forward-looking visionary leadership.

Although Bruce Chizen had worked at Adobe since the 1994 Aldus merger, he was very different from his predecessor. Unlike John Warnock, Bruce lacked a technical background, but as he often mentioned, his role was to inspire his engineers, rather than create products himself.

Although I always prefer technical leaders as CEOs of technology companies, Bruce Chizen’s successful 7 years as its CEO proved it could outlive its co-founders. During the period, Adobe began bundling its creative tools into one package, the Creative Suite, starting with Photoshop CS. Ever paranoid about competition as a company, Bruce acquired Macromedia, eliminating the threat of market share loss on the web as Macromedia’s Flash, used to create web videos and games, proved dominant, although short-lived.

During his seven years as CEO of Adobe, revenue grew by 14% annually while operating profits grew by 11% annually. Perhaps the even more successful management transition was from Bruce to Adobe’s current but now departing CEO, Shantanu Narayen, an Adobe veteran since 1996.

“I’m not leaving under any cloud. There’s no severance agreement. I’m taking a break, and the reason I feel comfortable taking a break is because the company is in such good shape” - Bruce Chizen, on leaving Adobe in 2007.

Test of time case study 3: Financial crisis

Beyond the technological disruption risk, Adobe’s other major business risk and challenge is its product demand sensitivity to the broader economic and business climate. Here’s an excerpt from its 2003 annual report:

“The Creative Professional market is sensitive to the economy, as most users of Creative Professional software products derive their revenue mainly from corporate marketing and ad spending in corporations. In difficult times, corporations tend to reduce marketing spending.” Adobe 2003 annual report.

23 years later, this challenge hasn’t changed and remains a problem, and in my risk analysis (section 8) I discuss more broadly how I account for this today.

Read more

Historically, investors relied on return on invested capital (ROIC) and free cash flow (FCF) margin, but as time has proven, there are many examples of companies that once had a high ROIC, only to decline a few years later due to industry cycles, limited pricing power and new competition due to low barriers to entry.

I’ve been thinking of ways to assess quality in a simplified way better and have created a 10-category low, medium, or high score quality table.

To explain how this works, I will utilise 10 of Warren Buffett’s famous investments and then discuss how the companies we have so far assessed on Global Outperformers perform on the same quality table.

The 10 quality attributes

Quality companies have many attributes, and while some are more important than others depending on the industry, geography or macro context, I believe the following ten attributes reflect most of what we look for in quality companies:

• Barrier to entry

• Switching cost

• Purchase frequency

• Mission-critical

• Technology resilience

• Value chain control

• Pricing power

• Balance sheet strength

• Profit margin

• Test of time track record

The quality table measures each of these attributes into three categories, low, medium or high and depending on the number of attributes in ‘low’, companies are either classified into:

• Ultra-high quality companies: These companies don’t have any attribute with a low score

• High quality companies: They have only 1 or 2 quality attributes with a low score

• Moderate quality companies: They have 3 quality attributes with a low score

• Weak quality companies: They have 4 quality attributes with a low score

• No quality companies: They have 5 or more quality attributes with a low score

The biggest limitation with this classification is that judging whether a company is low, medium or high still requires qualitative analysis.

The final column, growth, you’d notice in green, is similarly classified as low (<10% earnings growth potential), medium (10-15% earnings growth) and high (>15% earnings growth), and adds context for the quality table analysis.

To put this test into action, let’s next review 10 famous Buffett investments. On average, these ten companies returned 32% CAGR over the following five years after Buffett invested in them, with an average purchase P/E multiple of 11x.

• Ultra high quality (6/10 companies): Unsurprisingly for Warren Buffett, 6 of his 10 investments fell into the ultra high quality group with no single quality attribute with a low score. Buffett paid a higher than his average multiple for these companies at 15x P/E versus 8x P/E for the remaining 4 companies. On average, these 6 companies returned 34% CAGR in the five years after his investments. Finding these ultra-high quality companies at 15x P/E is extremely rare, and my further research showed that one could still achieve high returns paying 20-25x for these same companies, as long as they maintain medium growth prospects.

• High quality (2/10 companies): GEICO and Freddie Mac were both in the high quality category due to the shortcomings in pricing power and balance sheet strength for Freddie Mac. Note that Freddie Mac’s balance sheet weakness later impacted shareholders during the financial crisis. More on this shortly. These high quality companies still proved to be long-term winners, but compared to ultra high companies, Buffett paid much lower multiples of 8x P/E to account for their risks.

• Moderate quality (1/10 company): Wells Fargo was the only company classified as a moderate moat company with 3 quality attributes, switching cost, pricing power and balance sheet strength in the low score category. Buffett paid an even lower multiple of 7x P/E and demanded faster earnings growth from Wells Fargo to account for the quality shortcomings. Over the following ten years, Wells Fargo grew its net earnings at 26% CAGR, with its shares delivering an impressive 23% CAGR during the ten years.

• Weak quality (1/10 company): BYD was the only company in the weak quality category, and ironically, this was the only company Buffett himself didn’t directly invest in. At the time of the investment in BYD in 2008, it had a limited test of time track record, low profit margins, infrequent customer purchases for batteries and later electric vehicles and limited pricing power. While Buffett’s entry multiple of 11x was higher than what he paid for in moderate- and high-quality companies, he was compensated by BYD’s exceptional growth of 26% net profit CAGR over the 16-year holding period.

• No quality (0/10 company): None of the 10 Buffett companies was classified as a no quality company, and in line with his mantra of good companies at fair prices, he totally avoided these no quality companies. As followers of the Buffett approach, we should also follow this lesson for long-term investments

Some other key takeaways

1. Quality + value + growth: As we saw with the Buffett examples, going down the quality chain requires one to demand more value (lower earnings multiple) and growth (faster earnings growth) to make up for quality shortcomings. BYD was of much lower business quality than Coca-Cola, and in return, it had to grow even faster to make up for the quality gaps.

2. Ultra-high quality companies: These companies are extremely rare, and as we saw with the Buffett examples, when found with medium or high growth prospects- Coca-Cola grew at 13% CAGR for a decade, Washington Post grew at 20% CAGR for a decade- they have extremely high total return potential. It took Buffett 43 years to find 6 of these companies, averaging one new Ultra-high quality idea every 7 years. This is a reminder of the patience true high quality and long-term investing requires.

3. Quality score changes: During the 1970s, the Washington Post initially had a high technology resilience, pricing power and growth score, but by the 1990s, each of those attributes fell to low scores, and the return potential similarly declined. The lesson here is that we must continually assess quality and growth attributes to maintain attractive long-term returns.

4. Deal breakers: There were likely several potential investments Buffett considered but didn’t invest in because the low score in one of these 10 categories overrode any investment case. An example was his decision to sell Freddie Mac when he concluded he could no longer understand their balance sheet at the start of the 2000s. Sometimes, these low scores override any investment case, and like Buffett, we must think quite deeply about how bad a low score is when making any investment case.

Using the same framework, I have applied the quality table to the 15 companies we have so far deep dived on Global Outperformers and discuss them below.

• Ultra high quality (3/14 companies): Visa, Mastercard and International Container Terminal (ICTSI) each met the ultra high quality hurdle, but using the Buffett P/E standard, only ICTSI meets this as it was valued at 15x P/E at the time of our deep dive. Visa and Mastercard were each valued in the 28-30x P/E range, and it’s unlikely to expect those returns here.

• High quality (7/14 companies): Most of our deep dived companies fell into the high-quality category, but unlike Buffett’s investments, 5 of them have low scores in technology resilience. The Mexican airports, ASUR and OMAB, and FactSet also fall short on the growth prospects rating, with low scores here. Given we also paid higher multiples around 13-20x P/E for these companies, we also can’t expect the Buffett like returns here.

• Moderate moat (3/14 companies): Farmer Mac, Metlen and Latour each had 3 low quality attributes, and to account for their shortcomings, we first ensure there’s no technology resilience risk and pay only low P/E multiples of 9-10x for them. We should also include a growth hurdle of medium or high to ensure the low quality categories are well compensated for in the overall investment equation.

• Weak moat companies (1/14 company): Unsurprisingly, the worst performing company to date was also the only company with a weak moat score, Grupo Mateus. As we learned from Buffett’s examples, weak moat companies should be avoided in nearly all cases and only purchased when the growth prospects (BYD at 23% 16-year CAGR) are extremely attractive.

Concluding thoughts

There are several limitations to the quality table discussed above, and by no way should this replace thorough fundamental research. I view this table more as a reality check on potential investments to minimise potentially bad outcomes. For example, in the future, should a company have more than 2 low quality categories, it must have low valuations (<15x P/E) with medium or high growth potential if we are to expect attractive returns.

I plan to incorporate the lessons here even more, particularly the deal-breaker lesson discussed above when a low score might be too poor that it overrides any potential investment case. From past deep dives, both Grupo Mateus and Metlen are under further due diligence and monitoring from a balance sheet strength lens. As we will discuss in the next deep dive, Adobe could be in this category from a technology resilience assessment.

Register for the call

I’ll be hosting the Q2 2026 investor call later this week on Thursday 30th July, 5-6 pm (GMT).

The investor call discusses our most recent investment write-ups and portfolio changes. In Q2 2026, the focus was the software industry, and we published our research on Microsoft and Intuit, and have the Adobe deep dive coming up.

There will also be a Q&A, and you can either send your questions via the comment section below, in the registration form attached below or during the call.

To register for the virtual call, please fill out the form on the Jenga IP website here. A calendar link will be sent to you shortly.

Date: 30th July 2026

Time: 17:00 - 18:00 (GMT)

Location: Virtual

H1 2026 Performance Update

The 15 investment cases published on Global Outperformers since inception generated an equal-weighted average share-price return of 18.13% from their respective publication dates to 30 June 2026 (Q2/H1 2026), over an average holding period of 241 days.

These figures exclude dividends, currency movements and transaction costs and should not be interpreted as the return of a model portfolio. The average return remains ahead of the long-term annualised return required to achieve our 15% IRR objective.

The update reviews the principal detractors and lessons from the drawdowns experienced across the published cases.

Grupo Mateus and Metlen

Grupo Mateus, a Brazilian grocery chain and Metlen, a Greek industrial conglomerate, were the two worst performers, with each dealing with business-specific issues I had underestimated at the time of our initial research.

Metlen’s power project infrastructure, now folded into the energy division, experienced significant cost overruns in 2025, leading to a steep earnings miss in 2025. Cash flow generation remained weak, adding more risk to its business case. Management has since restructured the power project division and anticipates a turnaround in 2026 with its long-term €2 billion still in place. I’m carefully watching Metlen’s debt levels today as this could break its investment case.

Grupo Mateus experienced a more difficult than we assessed retail climate, leading to below expected growth projections across pricing and store count. Inventory miscalculations announced in Q4 2025 added to its string of challenges. Although we acknowledge its challenges, 6x net earnings doesn’t reflect its growth potential.

AI exposed (FactSet Research, Microsoft and Intuit)

The main theme we focused on during the first half of 2026 was asset-light software companies with recurring and profitable earnings that seem exposed to AI at first glance. Each of these companies experienced sharp drawdowns during the period, and trade near all-time valuation lows in the case of Intuit and FactSet. We are optimistic on each of these three companies and continue to view their long-term resilience as the investment case catalyst over the next five years.

ASUR B - Mexican airport group

ASUR has operationally and share price-wise underperformed its peers. Part of the underperformance was due to currency headwinds. The US dollar depreciated by 14% during the holding period against the Mexican peso. Given ASUR’s airports are more exposed to USD international fees and the Puerto Rican airport USD exposure, it broadly suffers during peso strength. Its largest airport, Cancun, has also underperformed traffic forecasts.

However, if we account for the currency appreciation and 15% dividend paid to ASUR shareholders last year, its real returns are positive, exceeding the 15% IRR hurdle rate to date. ASUR remains in our internal Jenga IP portfolio, and we view its current 12x forward earnings as attractive.

ICTSI and Alphabet

Both ICTSI, an emerging markets ports operator and Alphabet, the world’s largest search engine provider with other technology businesses, were among the companies (and TSMC) that returned more than 100% since our publication. While we continue to hold TSMC, we now view ICTSI and Alphabet valuations above levels we view as undervalued over the mid-term and have redeployed the proceeds to other undervalued companies discussed above.

Drawdowns

An interesting observation I recently noted on companies analysed to date is that among the 15 companies, 10 of them experienced a 15-20% (or more) share price drawdown within the first few months of researching and publishing their investment case. These drawdowns are a natural result of investing in companies unloved by markets, and over time, we’ve learned to take advantage of them. A good case study is Farmer Mac.

Despite achieving a 14% return in its first 6 months, Farmer Mac shares had declined by 21% within the first three months. Farmer Mac missed its Q4 2025 EPS estimates after reporting higher-than-expected provisions for loan losses for its agricultural loans. Some might panic or cut their losses in such moments, but after revisiting our deep dive and updating our analysis, we noted no material fundamental changes and took advantage of the decline, purchasing more shares at 7x its 2026 expected net profits.

Similarly, we recently experienced a 28% drawdown in OMA B driven by market worries on challenges from the ongoing Middle East tensions on broader air travel. While there’s undoubtedly some impact here, we certainly didn’t think it was worth a quarter of its market cap and similarly added to the position during the drawdown, at 13x its forward P/E.

The recent drawdowns will not be the last. Finding undervalued companies often means investing before market sentiment improves and temporary declines are therefore an unavoidable part of the process. Future investment cases will include more detail on position sizing and the valuation levels at which we would consider increasing our exposure.

Last week, I initiated (but now sold) a 5% position in Latour after it met my initial 15% IRR hurdle based on earnings estimates from my 2025 valuation model. However, I spent the past week poring through Latour’s filings and those of its 10 listed investments and concluded my initial assumptions from last year were too optimistic, and have now updated my Latour valuation model.

Based on the updated valuation, which I will walk you through in this report, I estimate that Latour’s shares offer an IRR of 11.3% to 2030. Given this is below my 15% hurdle rate for new investments, I sold my investment in Latour’s shares for a small 0.2% gain. Latour remains on my core portfolio shortlist, and after discussing its valuation, I will discuss what price would make me reenter its shares.

Latour's 2026 valuation model

A Latour reintroduction

Latour is a Swedish industrial conglomerate, part family office and part serial acquirer. You can read my extended deep dive here or a summarised version here. Its business model is split between its wholly-owned operations (7 companies) and its listed investment portfolio (10 companies). Let me first discuss the wholly-owned operations.

On average, Latour aims to achieve 10% revenue growth with 15% operating margins and return on operating capital over the economic cycle.

1. Wholly-owned operations

Latour fully owns 7 industrial companies. At their core, these diversified companies serve industrial niches, providing mission-critical tools and services to customers in the building and construction (Bemsiq, Innovalift, Nord-Lock, Swegon), warehousing (Caljan) and tradespeople (Hultafors) industries.

Although diversified and mission-critical, like most industrial companies, their earnings growth is still somewhat subject to the European and global industrial cycle. For example, if there’s a slowdown in building and industrial activity, tradespeople are left with lower work activity and incomes and are less likely to spend EUR 110 replacing their Snickers workwear trousers. Despite being critical for buildings, Swegon’s HVAC solutions demand will decrease if its residential property customers anticipate a real estate market slowdown.

Among these 7 companies, Swegon, a manufacturer of HVAC solutions, is the largest, representing 13% of Latour’s NAV or 28% of the wholly-owned division. Bemsiq (6% of Latour’s NAV), a serial acquirer in building automation and metering, has been the fastest grower of the 7 companies in recent years (22% EBIT 5-year EBIT CAGR), mainly driven by Latour’s aggressive acquisitions of smaller companies like Eelectron and QEL in the KNX home automation industry.

Finally, Nord-Lock, in the bolts and fasteners industry, is the most profitable of the 7 wholly-owned companies, often earning EBIT margins of ~25% and a return on operating capital of ~30%.

I will next walk you through their valuations.

Latour wholly-owned operations valuation

Despite being private companies, my valuation approach is still the same as with public companies. That is, estimating where earnings could grow to in a conservative base case scenario and concluding on an earnings multiple that best reflects its business quality (resilience, profitability, pricing power, etc.) and growth prospects. As a reality check tool, I also assess the earnings multiples of listed per group companies.

The table below summarises how I view Latour’s wholly-owned operations EBIT multiple, current valuation and earnings growth over the next five years and 2030 exit valuation.

The first two columns highlight how Latour values its holdings, which is on a price-to-EBIT multiple basis. Using this same approach, I revised my own exit multiple, which, on average, is just slightly less optimistic than Latour’s management, explaining why my valuation of its wholly-owned portfolio is roughly 2% lower or SEK 1.34 billion as at 2025 year-end (management’s SEK 64.6 billion versus Jenga IP SEK 63.3 billion).

Compared to management, I’m slightly less optimistic on Innovalift, Swegon and Latour Industries earnings growth and business quality prospects, which explains their lower exit multiple when compared to the median management range. For example, management values Innovalift at 16x EBIT while we value them at 14x EBIT.

The 2020 to 2025 five-year period was exceptionally impressive for Latour, with revenue and operating earnings growing by 13.3% and 14.2%, respectively, above its 10% annual growth target. It might be tempting to project that growth for the next five years, but the aggressive growth during this period came at the expense of balance sheet strength, with net debt more than tripling during the period (see chart below).

Given how aggressively Latour put its raised debt to work through acquisitions across its subsidiaries, it will be far more difficult to repeat such growth rates without exceeding its debt capacity limits. As a result, I expect acquisitions to play a lesser but important role for Latour over the next five years.

Finally, two of these seven companies, Innovalift and Bemsiq, were created during the period, and it’s typical for Latour to supercharge new spinouts during their first years. Bemsiq, for example, tripled its revenue and operating capital between 2020 and 2025.

It’s possible Latour could spin out or create a new platform company over the next five years, but to be conservative, I don’t include this in my analysis and view it as an added optionality to its investment case. To further assess how I concluded on these growth estimates, let’s briefly discuss their overarching themes.

Bemsiq and Innovalift - Aggressive acquisitions

As Latour’s newer companies, Bemsiq and Innovalift, have been more aggressive on the acquisition front and plan to maintain the pace with explicit goals of doubling 2024/2025 revenue by 2029. Although my own growth estimates are below management targets, between both companies, I believe Bemsiq has a higher chance of achieving this. Compared to Innovalift’s elevator services industry, Bemsiq’s metering and building automation offers more industry growth drivers and acquisition targets, given its less oligopolistic market structure.

From a profitability and return on operating capital lens, Bemsiq’s business also requires less capital, and will also contribute to its performance during the current currency headwind, tariffs, and broader geopolitical environment. Innovalift’s Turkey exposure through Arkel (est. at 19% of revenue) could add to near term challenges for Innovalift. That said, 40% of Innovalift’s revenue are from the more recurring components and maintenance division and I expect the division to standout.

Hultafors Group, Swegon and Nord-Lock - Latour’s stalwarts

Latour has held each of these three companies for over 20 years, and combined, they represent roughly two-thirds of Latour’s wholly-owned companies. Despite their maturity, revenue and earnings growth have remained impressive, with both compounding by 10-13% over the past five years, ahead of most publicly listed peers.

While it is tempting to simply project this growth for the future, I suspect Hultafors and Swegon will likely have a more difficult acquisition market while operating conditions could become more competitive. Swegon’s HVAC portfolio is more tilted to residential and office customers than growth markets like data centres, limiting its ability to capture mid-term growth trends.

For Hultafors, although PPE remains a significant share of revenue (69%), the hardware hand tools, particularly in North America, face a more difficult construction market, coupled with more pressure from lower-priced retailers Husky (Home Depot), Kobalt (Lowe’s) and Hyper Tough (Walmart).

From the table below, you might notice my increased bullishness for Nord-Lock’s revenue and operating earnings.

The more I study the fastener and precision-machined components market, the more I appreciate its attractive combination of consolidation opportunities, pricing power and exposure to structurally growing end markets. A recent example is Nord-Lock’s acquisition of the UK-based Energy Bolting, a manufacturer of specialist fasteners mainly catering to the oil and gas, power generation and marine industries.

Although its UK Companies House filings don’t reveal its income statements, its balance sheet is strong, and there are hundreds of companies worldwide that could fit into Nord-Lock’s portfolio.

“In 2025, sales exceeded SEK 2 billion, with profitability remaining strong. The ambition is to double revenue by 2028. This will be achieved through a combination of organic growth and selective acquisitions.” - Daniel Westberg, CEO of Nord-Lock Group

Caljan - Cyclical turnaround

As I had suspected last year, I expected 2025 to be a turnaround year after two years of declining sales and operating earnings, with the momentum continuing in Q1 2026. Caljan’s telescopic conveyors are quite lumpy in sales and dictated by e-commerce market cycles. During boom times, customers like DHL and Amazon are more likely to invest in additional conveyors, while during downtimes, there’s typically little demand for new machinery.

For the long term, I don’t see much growth for Caljan, and I expect them to surpass their 2022 peak earnings sometime between 2028 and 2030. 

Listed peer valuations

As a reality check on my valuations, I assessed listed peers, and the table below compares my valuation multiple to three listed peers for each Latour wholly-owned company. On average, my valuations are broadly in line with markets, which provides additional comfort that our estimates are conservative.

2. Latour’s listed portfolio

Rather than simply taking the current market cap of Latour’s 10 listed investments as the intrinsic valuation, I assessed these companies assuming we are making long-term investments in them, forming estimates of their 5-year revenue, EBIT and net income growth, and then concluding on P/E multiples that best reflect their business quality and growth characteristics.

While my individual analysis isn’t as deep as single company deep dives, I spent some time poring through filings, capital markets day transcripts, quarterly presentations, sell-side reports and broader materials on each of the companies to conclude on these figures.

The table above maps out my overall analysis of the ten listed companies, highlighting:

• Current 2025 end-of-year P/E

• My 2030 exit multiple

• 2025 - 2030 revenue, EBIT and net income CAGR estimates

• Current valuation of Latour’s stake

• 2030 Valuation of Latour’s stake based on the exit multiple and net income estimate

As concluded from the table, I expect Latour’s publicly listed portfolio to deliver a 5-year annualised return excluding dividends of just 2.9% (SEK 88.7 billion to SEK 102 billion by 2030), and three problems contribute to my lower return estimate:

• Assa Abloy and Securitas growth challenges

• HMS Network and TOMRA - Expensive valuation

• Sweco - A more challenging AI reality

Before discussing these three issues, I want to highlight management’s own guidance from their Capital Markets Days.

Excluding HMS Network’s 15% revenue growth target and possibly Assa Abloy’s 10% revenue growth target, I suspect it will be challenging for the rest to meet these management revenue targets over the mid-term.

On a margin basis, the management targets appear more achievable. Although my estimates differ slightly (see valuation model), they are roughly in line with my own margin estimates during the period.

Assa Abloy and Securitas - Growth challenges

Both companies are Latour’s largest (Assa Abloy - 27%) and 5th largest (Securitas - 7%) holdings, and combined, represent a third of Latour and, as such, their outcomes severely impact Latour. Historically, both companies have done fairly well, but as they’ve gotten larger, it’s harder for Assa Abloy’s acquisitions to drive incremental growth while Securitas’s larger scale poses threats to labour costs and global wage inflation.

To date, Assa Abloy has made just around 400 acquisitions and a decade ago, they could probably achieve their 5% annual acquisition growth with just 10 companies, but today, it needs 25-30 acquisitions to meet the same target. In Q1 2026, Assa Abloy completed just 3 acquisitions, representing 0.4% of its 2025 revenue.

Encouragingly, both companies are managing their costs well and have so far dealt with tariff challenges well, with Assa Abloy raising prices across its products.

HMS Network and Tomra - Expensive valuation

If valuations didn’t matter, both HMS Network and TOMRA, two fairly profitable, acquisitive, and organically growing companies, could have joined our portfolio individually. The challenge, though, is that valuations do matter, and at the start of the year, they were valued by markets at 48x and 40x their respective net profits.

At such high valuations, both companies would have to grow profits north of +25% CAGR to justify such valuations while making an attractive IRR. The cost of this high starting valuation is already affecting Latour with Tomra’s shares declining -38% YTD after announcing a Q1 2026 profit warning.

Sweco - A more challenging AI reality

I marginally reduced my exit multiple from 23x to 22x for the Swedish architecture and engineering consultancy due to AI. While AI doesn’t replace architects, as I study how companies are implementing AI, I see a world where, rather than requesting 20 billable hours, Sweco’s customers could build initial drafts with AI and then request 15 billable hours, thus reducing total hours required.

Compared to more traditional technology or management consultancies, Sweco is certainly more insulated from this risk, but not fully immune. I suspect this risk partially contributed to the slowed sales experienced in Q1 2026, and I have revised my revenue and net income estimates. To combat this risk, though, I take comfort in Sweco’s above-inflation revenue per employee across its 8 regions, particularly in the UK, Germany and Central Europe.

At current market prices, Sweco still offers the most attractive IRR among the Latour listed investments and I have upgraded them to my core shortlist.

Latour’s valuation

Latour's 2026 valuation model

Based on the wholly-owned and listed investments analysis, I arrived at a Latour valuation model by mapping out the net debt (see grey bar above) and annual dividend payments, which is guided by its target of paying 100% of dividends received from listed companies and 50% from private companies out to shareholders. 

Overall, this leads to a 2030 estimated NAV of SEK 177.4 billion or SEK 277 per share.

Based on current valuations, this leads to an IRR of 11.26%, above 9% for an undervalued status but below the 15% IRR target for new investments which led to my sale.

Should Latour’s shares decline to SEK 161 per share, I’m happy to reenter its shares. But for now, Latour stays on the core watchlist.

Last week, I published a deep dive on Intuit. Below is the abridged one-page summary of its investment case.

Company overview

Read more

A PDF version of this Intuit deep dive is available to premium readers at the end of the report.

I have an investment proposal for the next 30 years. During the holding period, here’s what I guarantee will happen:

• Its flagship product will become irrelevant within a decade

• Operating profits will decline for the first five years

• Microsoft will spend the next decade attacking its main products

• It will face a leadership crisis once its founder steps back

• Three technology shifts (internet, mobile and AI) will challenge its competitive advantage

• Lastly, you can expect at least three 50% share price drawdowns during the holding period

Most rational and risk-conscious investors would likely pass.

Despite these challenges over the past three decades, Intuit’s shares delivered a 42-fold return, or 13.4% annualised return, quadrupling the S&P 500’s return (11-fold or 8.4% annualised return) during the same period.

Like previous technological shifts over the past three decades, there’s once again doubt and worry about the future of the five foundational software companies (Adobe, Intuit, SAP, Oracle and Microsoft)

Market capitalisation (As of 21st May 2026): $84.9 billion

Jenga IP 2030 FY estimated market cap: $161.9 billion

Potential IRR (including dividends and buybacks): 20.3%

Jenga IP Quality Rating: 76.7/100

“Quicken [Intuit] is over! … It’s done. It’s almost a non-factor. Intuit is learning the hard way that owning a market niche doesn’t mean much when new technology comes along and washes it all away” - David Farina, August 1997, Fortune issue.

Among the five companies, Intuit arguably had the most fragile-looking path and now faces existential AI-era questions. Its investment case now sits on five questions:

• Can Intuit maintain its ecosystem control of financial services to American SMBs?

• Can Intuit reduce churn across its payroll and payment solutions to SMBs?

• Does TurboTax become a higher moat business as it transitions from DIY to expert assisted?

• Does AI serve as a net positive or negative for Intuit’s competitive advantages?

• How lucrative do OpenAI and Anthropic perceive the bookkeeping and tax preparation market?

To answer these core questions, here’s what I’ll cover in the Intuit deep dive:

Table of contents

• Intuit’s history: Insights into Intuit’s transition from personal finance (Quicken) to a multi-product SMB and consumer ecosystem, efforts in scaling SMB and consumer software solutions across the decades, case studies on product challenges and pivots, culture and capital allocation, and the durability and weaknesses of its competitive advantages.

2. SMB financial services industry: An overview and analysis of Intuit’s SMB financial services markets, ranging from accounting software to payroll and payments. A comparison between Intuit’s QuickBooks and key competitors (Xero, Sage, Block and Paychex). The customer’s view across the U.S. and international markets.

3. The tax preparation industry: The legal, historical and competitive architecture of U.S. tax filing, TurboTax’s market position, assisted versus digital shifts, IRS Direct File and the long-term durability.

4. Intuit’s product suite: An assessment of Intuit’s product scope (QuickBooks, Mailchimp, TurboTax and Credit Karma), product plans and pricing, product architecture on a usage and per seat monetisation and usage versus per seat economics and product innovation.

5. Intuit’s AI opportunity and challenges: A deep dive into Intuit’s AI efforts, customer feedback and traction, an overview of AI-native challengers targeting Intuit’s solutions and an analysis of solutions at risk from AI.

6. Intuit’s business economics: Analysis of Intuit’s revenue and costs by segment (Global Business Solutions, Consumer, Credit Karma and ProTax, economic contribution of acquisitions such as Mailchimp, broader assessment of Intuit’s competitiveness and a unit economics analysis of QuickBooks and TurboTax.

7. Growth opportunity: Analysis of its growth drivers via ARPU expansion, customer growth and insights into its potential product expansion, a review of past international growth failures and current opportunities.

8. Risk and challenges: Further assessment of the competitive threat and challenges across AI-native, payroll specialists and vertical software companies, overpricing risks, tax simplification attempts, share-based compensation reliance and the non-GAAP reconciliation, and final thoughts on AI-driven seat compression and ecosystem control.

9. Valuation: An earnings multiples-led valuation and projection of Intuit’s earnings potential through to 2030 FY, comparison with peers and the broader software industry.

10. Conclusion: Final thoughts on the software downturn, insights into portfolio management and sizing a potential investment in Intuit.

Subscribe now

1. Intuit’s history

Like many software companies, there’s some scepticism about Intuit, a provider of bookkeeping, tax and personal finance solutions, in today’s AI world. Will people file taxes directly on LLMs? Will businesses transition their payroll and accounting directly to chatbots rather than Intuit’s QuickBooks? Or will individuals depend more on AI platforms for personal finance planning than Intuit’s Credit Karma?

To answer these questions, it’s important first to understand where Intuit has come from, why it became the most successful bookkeeping platform globally despite 46 other software companies coming before them, and why larger technology companies like Microsoft failed to beat Intuit in its own game.

To appreciate its business resilience and anti-fragility to date, it’s best to examine its evolution in four phases:

• Phase 1 1983 - 2000. The Scott Cook era

• Phase 2 2000 - 2008. The Steve Bennett era

• Phase 3 2008 - 2018. The Brad Smith era

• Phase 4 2018 - present. The Sasan Goodarzi era

Phase 1 1983 - 2000. The Scott Cook era

For the first decade, though, Intuit wasn’t in any of these markets. 3 of its 4 core platforms were acquired while QuickBooks grew out of what Intuit was initially known for, Quicken, or as initially called, Kwik-Chek.

Intuit’s founder and current board member, Scott Cook, noticed his wife’s challenges with balancing their family’s checkbook, and he figured he could solve this challenge through personal computers. Although his marketing experience at Procter & Gamble proved valuable in shaping Intuit’s growth strategy over time, to get off the ground, he lacked the required technical software development knowledge, and to fill this gap, Intuit’s co-founder, Tom Proulx, a then student at Stanford University, joined as the programmer and wrote the first version of Quicken.

Although certainly important, Quicken’s success wasn’t simply due to outcompeting peers in technology. There were already 46 players in the personal bookkeeping market prior to Quicken’s entry, and their success was mainly due to four reasons:

• Simplicity: Inspired by Apple’s Lisa user interface, Scott Cook concluded that the best way to approach the personal finance software market, at that time, was ensuring its user interface looked like a real-world checkbook interface. Then market leaders like The Home Accountant and Managing Your Money by MECA had focused too much on features and showcasing their software capabilities rather than actual customer needs.

  Financier Personal Series by Financier Inc of Westboro, Mass had even gone as far as enabling 33,000 types of accounts, with its listing price set at $195, more than double Quicken’s price. Quicken, unlike peers, was fully focused on actual customer needs guided by customer research. The philosophy here inspired the company name “Intuit” as the tools needed to be “Intuitive”.

“Manage Your Money by MECA Software of Trumbull, Connecticut, was so complicated that the staff at a software convention MECA booth failed to make it work. “This program is like an adventure game, it’s so hard!” A MECA employee exclaimed to Cook.” Inside Intuit by Suzanne Taylor

• Follow Me Home: Scott Cook’s consumer background was particularly helpful for Quicken in winning customers. To ensure the experience left customers “delighted”, Scott Cook instituted a “Follow Me Home” research process where both marketing and engineering staff followed willing first-time customers home from software stores, watching them install and use Quicken.

  While closely watching customers, engineers took notes of any feature lags and when customers seemed confused, hesitated or reached for the manual for further explanation. The feedback here, coupled with the internal database order-entry system created by Proulx to track customer comments, ensured that Quicken was optimised for the customer experience and user ease.

  Every employee in the company answered customer service and support calls at least four hours each month and also created a customer advisory panel, led by loyal Quicken customers. Today, these seem obvious from a business strategy, but in the wake of the software world in the 1980s, customer obsession wasn’t central to business strategy.

• Frugality: A key reason Intuit, alongside other foundational five software companies like Microsoft, SAP and Adobe, survived and thrived for decades was their frugality due to limited funding and operational budgets. Despite being right at the heart of Silicon Valley, Scott Cook only raised $151,000 from friends and family to launch and scale Intuit.

  Frugality was necessary for survival and guided every decision at Intuit. For example, to cut COGS by half, Quicken resized the program disk and manual to save on packaging.

  After partnering with banks to sell Quicken to customers, management streamlined marketing and advertising to the beginning or end of their fiscal years, saving unnecessary costs. By the fourth quarter of 1985, Intuit became cash flow positive.

• Partnerships and distribution: Given the limited budget, Intuit couldn’t initially scale via retail campaigns and stores as aggressively as larger and more established peers did. As alternatives, they focused on working with partners who benefited from Quicken’s rise.

  For example, Apple admired Quicken’s interface and ease, and it also increased the use case of Apple’s personal computers. Banks across the U.S also benefited from Quicken because it made customers more engaged with their bank accounts, driving account retention and product purchases. One of the successes here was the Bank of Hawaii, which single-handedly drove Quicken into becoming the top consumer software in the state.

  The financial success gained from these partnerships provided Intuit with the resources to then expand through the more traditional software stores, signing deals it couldn’t previously close due to its then small scale. Between 1988 and 1990, sales grew five-fold, from $6 million to $33 million.

You may have also noticed that unlike other foundational five software companies, the product name (Quicken) and company name (Intuit) were different from day one. This is quite important because Intuit has never been about a single product or solution. It’s always been about simplifying financial management for customers, and Quicken was simply a stepping stone for a longer journey.

Always on the hunt for expansion, Intuit began surveying small businesses as early as 1988 to gauge if it could broaden its tools beyond personal finance. It was immediately evident to Cook there was an opportunity among SMBs, and he later hired Sam Klepper to lead the efforts, and after noticing a Quicken Payroll add-on called QuickPay, created by Mike Potter, Intuit doubled down on separating the SMB accounting service, and QuickBooks was launched in April 1992.

Unlike consumers, SMBs couldn’t avoid periodically updating their bookkeeping, making QuickBooks more of a software staple than Quicken’s discretionary use case (another key quality attribute), which led to a more recurring, stickier and more resilient to economic shocks product base. By the second half of the 1990s, it became clear to management that QuickBooks would be more important for Intuit than its original Quicken, and they would later sell Quicken in 2016 to focus on QuickBooks.

Acquisitions also played an important role in growth, although not all of them worked. In 1993, it acquired ChipSoft, owner of TurboTax, further increasing the business stickiness, while bolt-on acquisitions like National Payment Clearinghouse in 1993, Parsons Technology in 1994 for SMB software, GALT Technologies in mutual fund distribution and Lacerte software in 1998, addressing the professional tax software, supported the core Quicken and QuickBooks platforms.

Intuit had also diversified into physical financial supplies like printed checks, deposit slips, tax forms and payroll checks and by FY 1997, they reached 15% of their revenue, growing even faster than pure software sales.

Phase 2 2000 - 2008. The Steve Bennett era

Despite the impressive revenue growth, from $33 million in FY 1990 to $1,037 million by FY 2000, a 41% CAGR and share price return of 645% between its IPO and the end of 2000, Intuit still wasn’t a rock solid company. Markets worried about the health of Quicken given the cyclical and fad nature of personal finance software (more on this when we discuss Credit Karma later on), and between Nov 1995 and April 1997, Intuit’s shares lost 72% of their value as profitability still dwindled. To improve its fortunes, Intuit promoted Bill Harris, who had joined through the ChipSoft acquisition (parent of TurboTax), to lead Intuit’s internet aspirations.

During his tenure, not only did profit margins remain tepid with slowing sales growth, there were culture clashes between Harris’ management style and the Intuit team, leading to his shortly lived CEO tenure. To ensure the board didn’t repeat this mistake, Scott Cook took a more hands-on role in picking the next successor and Steve Bennett was hired from GE in January 2000 to turn around Intuit’s fortunes and refocus the group into productive and profitable segments.

“He [Bill Harris] took the job around August 1 and within a week he was floating around book ideas. He was supposed to be the CEO!” Inside Intuit by Suzanne Taylor

Bennett’s more decisive and bottom-line focus, the Jack Welch way, was required during this period as the tech bubble burst. The increased focus was most evident in the QuickBooks 2003 release, as the bug rate had reduced by 70%, driven by process excellence techniques and defining specific tasks by phases.

“At one meeting, some managers presented a forecast for increasing page views. He said let me understand something: if you get more page views do you get more revenues?… So what drives revenue? So if page views go down you can still get more revenues.

“Steve taught us you can have hundreds of measures but if you’re not measuring the right things or if they’re too hard to measure you can really mess up.”

Coupled with improvements in TurboTax, which had reached a market share of 71% by 2003 with over 5 million desktop sales, Intuit’s financial performance greatly improved during its GE era of rightsizing and efficiency, with operating profits growing from $10 million in FY 2000 to $644 million by FY 2008.

Commentators often overlook this era of Intuit due to its share price decline of -7% during this period, but this was really the most important period for Intuit’s journey in transitioning into a financial management operating system.

Case study: The Financial Crisis

Although Intuit is very different from where it was during the financial crisis, it’s always useful to examine how companies deal with crises. Another pillar of my quality investment thesis for Intuit is its ability to withstand challenges.

Crisis or not, individuals still need to file taxes (+7% FY 2009 growth), while businesses still demand bookkeeping (-2% FY 2009 growth), file payrolls (+8% FY 2009 growth), and make payments (+14% FY 2009 growth). On the other hand, personal finance, being more discretionary, often experiences less user demand, and its revenue fell by 14% in FY 2009.

Intuit maintained group-level operating profit growth, supported by acquisitions and while certainly not a Visa or Microsoft in terms of resilience to downturns, Intuit performed much better than many other companies supporting the financial economy and those providing software to small businesses.

Beyond dealing with the crisis, Intuit, like many legacy software companies, was also dealing with the technology transition to mobile computing and the cloud, setting up for the Brad Smith era.

Phase 3 2008 - 2018. The Brad Smith era

Prior to his CEO promotion, Brad Smith initially led the accountants division at Intuit before leading both the consumer tax and small business (now called global business solutions) divisions, driving both revenue and market share growth during the five years.

If Steve Bennett brought stability, Brad Smith brought a growth culture to the group. Intuit’s challenge post its GE-era years was that it lacked innovation and the ability to scale new products to customers. Only 4 of the 50 products it launched in the prior decade grew above $50 million. Brad Smith immediately identified this, transitioning Intuit from a top-down R&D approach to bottom-up with smaller teams, sometimes as little as just two people.

Intuit was also among the faster-moving legacy software companies with the transition to the cloud and mobile computing. QuickBooks users had historically upgraded every other year, but the cloud’s economics of transitioning QuickBooks to a more recurring monthly model incentivised Intuit during this period. For example, QuickBooks Online Plus was initially available for $34.95 per month, while Payroll for QuickBooks Online was priced at $19.95 per month.

“Usability is the hallmark of Intuit’s desktop accounting products, and the online version carries on that tradition. Overall, the design is exceptional; I like it better than what comes out of the box.” - PCMag.com Product review

The Brad Smith era was quite successful, with Intuit’s EBIT growing from $644 million to $1,560 million (9% CAGR), with its shares rising by 582% during the ten years.

Although market commentators painted the decade as a completely rosy period, Intuit still had some failures. Intuit had ignored the international market, focusing mainly on the U.S., which created an opportunity for Xero to win share and eventually leadership in key markets like the UK and Australia, among others.

Quicken continued to decline and struggle in the cloud world, leading to an eventual sale in 2016. Several products like SnapTax, QuickBase (low-code workflow builder), Intuit Health (bought in 2010, sold in 2019) and Intuit Homestead (website creation and hosting) were shut down. Despite these failures, Intuit still delivered profitability growth as progress across QuickBooks and TurboTax mattered more.

Phase 4 2018 - present. The Sasan Goodarzi era

After 10 years, the CEO role moved to Intuit’s current CEO, Sasan Goodarzi, a then 14-year Intuit veteran, having worked and led both the consumer tax and SMB units. Sasan inherited a very healthy Intuit, centred between both QuickBooks and TurboTax, and his goal was transitioning Intuit away from an “SMB + tax” into the financial operating system. Achieving this goal relied on two strategies: the acquisition path and the artificial and human intelligence strategy.

Acquisitions

Intuit has always had a soft spot for the consumer finance market, and to reenter the market after its sale of Quicken, Intuit acquired Credit Karma for $8.1 billion (10x revenue) in 2020. Unlike Quicken, Credit Karma is less focused on the bookkeeping aspect of consumer financing; rather, it’s more focused on the “lending marketplace” aspect and is somewhat diversified across credit cards, personal loans, insurance and auto loans. We will get deeper into its business and growth prospects, but overall, it’s clear Intuit overpaid for Credit Karma.

The second major acquisition was Mailchimp, and likely an even costlier mistake was acquiring the customer marketing software business for $12.1 billion (15x revenue) in 2021. Prior to Mailchimp, Intuit had various attempts in this market. Demandforce, acquired in 2012 but later sold two years later, supported SMB customer marketing and automation. The Homestead acquisition in 2007, which became Intuit Websites, addressed some of Mailchimp’s current services that struggled in the face of Wix and Shopify.

Artificial and human intelligence

The more successful strategy in the Sasan Goodarzi era has been merging both artificial and human intelligence for existing tools and solutions. On the human and tax side, TurboTax Live, initially launched in 2018, connected users with tax experts, filling a gap its chief rival, H&R Block, had over Intuit.

A year later, Intuit launched QuickBooks Live, similarly connecting SMBs with bookkeeping and accounting experts with a range of services. These solutions increased the network effects, customer experience and value chain relationships for Intuit and today, both continue to grow quite fast, strengthening Intuit’s overall moat.

To further deepen the financial ecosystem, there’s also been further prioritisation on cross-selling across the various products;

• TurboTax and Credit Karma: Credit Karma users can directly file taxes through TurboTax in the Credit Karma app while TurboTax users have access to Credit Karma recommendations, driving commission and ad revenue. To further deepen this, Intuit officially merged both units in FY 2026 as the consumer segment.

• QuickBooks and Mailchimp: Given both serve SMBs, bookkeeping QuickBooks users are often sent emails on ways Mailchimp could amplify their customer reach, and QuickBooks open architecture directly syncs both applications together.

To further strengthen this ecosystem model, Intuit’s culture under Sasan Goodarzi has also transitioned more to the open system architecture, similar to other larger and modern technology companies. Intuit discontinued bets it concluded it couldn’t win, integrating deeper with companies like Shopify, Square, Stripe and other AI-native apps.

Until the past year, the Sasan Goodarzi era had been successful from a revenue and operating profit growth and share price performance view, with earnings compounding by 17.9% CAGR and a share price return of 304% (24% CAGR); that was until its most recent 50% share price decline, which brings us to where we are today.

Wall Street has voiced its concerns for Intuit’s future in today’s AI-driven world, and as we have learned from reviewing its history, this isn’t the first or second time markets have been worried about Intuit’s future. It’s also dealt with several technological shifts. Despite initial struggles and failures at each, from the internet during Scott Cook’s latter years in the 1990s to Brad Smith’s mobile computing challenges in the early 2010s, Intuit survived each, and I believe this is due to how embedded it’s become for SMBs.

Before exploring its product suite and economics, let’s first discuss the markets Intuit operates in.

2. SMB Financial Services Industry

Intuit is a technological conglomerate operating in several vertical markets, and to best understand its competitive landscape and wider industry prospects, it’s best to break the industry analysis across the main areas. Within the financial services segment, Intuit operates in four verticals: accounting and bookkeeping software (QuickBooks), payroll, payments and customer marketing and outreach (Mailchimp).

Let’s first look at the largest of these four markets, the accounting and bookkeeping software industry.

Accounting and bookkeeping software

Read more

A year ago, I made the investment case for several transport infrastructure companies around the world: airports, ports, railroads and highways. Unglamorous industries that make quite boring deep dives.

I first discussed the airport business model and then laid out the investment case for 2 of 3 Mexican airports (OMA and ASUR).

A month later, in June 2025, I discussed why I was bullish on International Container Terminal, a Philippines port operator and also discussed why Airports of Thailand was ripe for a rebound in my mid-year Global Equities report.

Finally, I curated a list of 10 transport infrastructure companies in the final reflections article that I thought were high quality and attractive valuations.

So have they since performed?

Not too bad!

An equally weighted portfolio of the ten companies I recommended returned +27.1% (or +29.2% with dividends included), marginally above both the S&P 500 (+25.2%) and the MSCI World Index (+24%).

Interestingly, just before the Strait of Hormuz crisis in mid-February earlier this year, the transport infrastructure portfolio was roughly 20 percentage points (37.4% versus 17%) above the market benchmarks.

This is a reminder of the inherent risks in transport infrastructure. Despite being high barrier to entry assets, disruptions to the flow of goods and people directly impact their valuation sentiment and fundamentals.

Returns contributors

• Class A (good): The portfolio was led by two significant contributors, ICTSI (+94%) and Airports of Thailand (+92%).

• Class B (average): The rest of the returns were more muted. Three produced satisfactory returns: Asian Terminals (+21%) (now gone private) and the two North American railroad companies, Union Pacific (+19%) and Canadian National Railway Company (+14%).

• Class C (bad): There were two negative performing companies, ASUR B (-9%) and OMA B (-6%), while three returned single digits; Aena (7%), Athens International Airport (+3%) and GAP B (+3%).

As you might notice, the majority of the transport infrastructure gains came from two companies, while the losses weren’t too severe, thus ensuring the whole portfolio remained resilient.

Beyond the 10 companies I shortlisted, there are 197 other listed transport infrastructure companies (GICS definition) above a market cap of $50 million. While we identified two top winners, there are a couple of others worth highlighting.

1. Emerging and frontier winners

One of the conclusions we had from the deep dives last year was that transport infrastructure is one of the very few industries in which emerging and frontier markets do better than developed markets. This is mainly due to the natural currency hedge and pricing power potential in emerging and frontier markets when compared to the more regulated developed markets.

• Nigeria - Sky Aviation (+160%) and Nigerian Aviation (+149%)

I am embarrassed to mention that two of the top performers were Nigerian airport logistics companies. Despite being Nigerian, travelling there earlier last year, meeting management at both companies, and even discussing both in the same article where I shortlisted the 10 companies, I missed both companies and shareholders more than tripled their return in USD terms.

• Jamaica - Transjamaican Highway (+113%)

Transjamaican Highway (+113%), a concessionaire and operator of two toll roads in Jamaica, was another big winner. They were also on my radar during the year and in fact, selected them as the best listed toll road opportunity I had come across in markets (7x P/E, 65% EBIT margins with 10-15% operating profits CAGR potential).

My worries about the long-term political threat to toll roads’ pricing power proved irrelevant in this situation, and growth remained strong, particularly in the May Pen and Spanish Town Toll Plazas.

Since going public 6 years ago, Transjamaican Highways has now been a 6-bagger, or 9-bagger with dividends reinvested.

2. Fraport AG (+20%) - The reader’s recommendation

A couple of readers brought to my attention that I had ignored Fraport’s turnaround opportunity when I first published the deep dives. They were right.

Fraport (+20%) was the best performer among the 7 major European airport stocks, beating my own European recommendations, Aena (+7%) and Athens International Airport (+5%).

Unlike the emerging markets companies discussed earlier, I haven’t followed Fraport as closely and will spend the next few weeks closely analysing its fundamentals.

Betting on transport infrastructure

There were multiple other winners around the world, and the year proved once again that transport infrastructure remains an industry that an investor with a global equities and quality focus should follow closely, given the breadth of investment opportunities available.

From an allocation standpoint and as I’ve spent more time learning the ins and outs of the industry, I’ve increased my maximum portfolio weighting from 15% to 20%. I also remain optimistic about the companies we deep dived last year (OMA, ASUR, ICTSI) and plan on updating their investment cases to premium readers in the coming weeks.

Current Middle East tensions have also improved the valuation entry points to new potential names, and I have included Aena in the core portfolio watchlist. Should valuations further drop, you can expect an extensive deep dive in the coming months.

TSMC’s shares have doubled since I published my investment case deep dive last June, surpassing my 2029 estimated share price and IRR target. Despite this, I remain long TSMC and have since updated my valuation model.

In this article, I’ll walk you through the rationale behind the changes I made to its revenue, earnings and free cash flow estimates for the next five years, alongside an updated valuation spreadsheet behind the paywall.

TSMC has historically been, and remains, a cyclical company with long-term secular growth through the semiconductor cycle. I reviewed earnings calls over the past several years, and the table below highlights the 13 times management provided forward revenue or profit guidance, and as shown below, they’ve been historically conservative, only failing to meet their guidance on two occasions in the last 15 years.

TSMC’s long-term growth target increased from 10 to 15% CAGR between 2020 and 2025 to 25% CAGR between 2024 and 2029, and the key growth driver is the ongoing AI boom, a theme I discussed during my initial deep dive but certainly underappreciated when I reflect on their customers’ updated growth ambitions, like Nvidia and Broadcom and their customers’ customer, such as Microsoft and Alphabet.

Like all contract manufacturers, TSMC responds to its customers’ demand, rather than creating demand, and by ensuring it has the most cost-effective and cutting-edge technologies available to customers, it can maintain market share and unit economics lead over rivals. There have been some updates since the TSMC deep dive, and I’ll discuss some of these across four topics.

1. The foundry market

There are different ways of calculating the global foundry market, and the two main methodologies I focus on are TSMC’s own reported foundry 2.0 market, which combines global foundry production and the advanced packaging industry, and the more direct revenue approach, comparing the revenue figures across the six largest publicly listed foundries.

TSMC reported the foundry 2.0 market reached $305 billion, growing by 16%, less than half of TSMC’s own 37% 2025 revenue growth. It’s quite astonishing for a market leader to achieve this, and 2025 was another year TSMC maintained a 15 percentage point growth gap to the foundry 2.0 market and now represents just under 40% of the whole market.

From a more competitive lens, TSMC’s revenue share among the six major listed foundries increased from 61% to 67% in 2025, achieving three times as much revenue as Intel and Samsung foundries combined.

It might be tempting to project this growth gap between TSMC and its peers for years to come but as we discussed during the investment case, 2025 was the perfect storm for TSMC because the latest 3-nanometer process node, being the last major TSMC node based on FinFET transistor architecture meant the ROI for other leading-edge foundries like Samsung and Intel expanding production and R&D was much less than prior years.

“Today’s technology is so complicated. So, once you want to design a very complete or advanced technology, it takes 2 to 3 years to fully utilise that technology. After 2 to 3 years of preparation, you can design your product. Once you get your product approved, it takes another 1 to 2 years to ramp it up” - Q4 2025 TSMC earnings call. Che-Chia Wei, CEO of TSMC

2. Leading-Edge process nodes

TSMC’s most leading-edge process node historically reaches 25% of TSMC’s wafer revenue by the third year and in 2025, the 3-nanometer node came in just under the historical hurdle at 24.3% and while it might be tempting to say it has so far underperformed, I think the more important takeaway is the staying power of the 5-nanometer process nodes, now in its sixth year, growing 37% year on year. For context, the 7-nanometer node revenue fell 33% in its sixth year (2023) and was already much smaller than the 5-nanometer node.

While TSMC doesn’t reveal its process node revenue by customers, I suspect its AI accelerator chips, the data centre custom chips and Nvidia’s H200 (currently utilising the 5-nanometer process node, TSMC N4) continue to drive its demand.

Although my 2025 wafer revenue estimates weren’t too far off from TSMC’s actual performance, I underestimated the staying power of both the 5-nanometer and 7-nanometer process nodes, as their wafer revenue was ahead of my estimates. For 2026, management discussed how they’re optimising capacity as they prepare for more demand in the 3-nanometer process nodes relative to both 5 and 7-nanometer:

“In addition to all the new fabs, we continue to convert 5-nanometer tool to support 3-nanometer capacity in Taiwan” - Q1 2026 TSMC earnings call. Che-Chia Wei, CEO of TSMC

Another interesting update was on the future process nodes, the 2-nanometer and A14 process nodes. The 2-nanometer node introduces a fundamentally new technology architecture, and both Intel and Samsung are more focused on achieving market share growth in this process node, over the current 3-nanometer nodes.

Earlier today, Apple, TSMC’s second-largest customer, was rumoured to be in preliminary talks with Intel and Samsung, which deviates from the exclusive partnership it had with TSMC from its iPhone 8 (2017) to the iPhone 17 generation.

Since my initial deep dive, the two leading-edge competitors, Intel and Samsung, are also in a much better financial position, publishing better financial outlooks than I imagined. For example, the memory chip market shortage has worked in Samsung’s favour, with the division representing 90% of Samsung’s operating profit growth in 2025. The additional profits here can be reallocated to supporting investments in its logic foundry business, leading to a more competitive market than we saw in 2025.

3. Capacity and fab expansion

The second biggest growth driver for TSMC is its capacity expansion, particularly across its most advanced chips, and capacity continues to track well, with new fabs across Arizona, Taiwan, Japan and Germany in line with the schedule. In Arizona, TSMC mentioned it is speeding up production due to customer demand:

“Construction of our second fab is already complete and tool moving and installation is planned in 2026. Due to the strong demand from our customers, we are also putting forward the production schedule and now expect to enter high-volume manufacturing in the second half of 2027” - Q4 2025 TSMC earnings call. Che-Chia Wei, CEO of TSMC

Although the manufacturing scale in Arizona would likely never reach Taiwan’s scale, the fabs here address a major long-term supply chain risk, bringing U.S. chip production onshore, and if TSMC can meet its goal of 30% of 2-nanometer capacity in Arizona, I believe this levels the political narrative relative to leading-edge peers.

Despite its international buildout, TSMC isn’t slowing its domestic capacity growth ambitions and is now preparing for additional 2-nanometer capacity in both Hsinchu and Kaohsiung Science Parks and an A14 processing facility in Taichung (Fab 25) scheduled for H2 2028.

Altogether, it’s clear that TSMC is preparing for even more demand than anticipated in 2024, and I have also increased my wafer capacity and shipment forecasts from 2026 through 2030 (see table below), and expect TSMC to add twice as much capacity in the next 5 years as it did between 2020 and 2025, with utilisation steadily above 95% for the next few years.

4. AI-led growth

Although it’s still unclear if AI will make TSMC more or less cyclical in the long term when compared to the smartphone era, it’s clear that the growth today is overwhelmingly driven by AI and data centre demand, with the High-Performance Computing (HPC) division reaching 58% of TSMC revenue, from 37% four years ago. Discussing the strong demand outlook, management once again reconfirmed that they expect mid-to-high 50s CAGR between 2024 and 2029 in their AI accelerator business and that it will be the biggest contributor of incremental demand across all its platforms. Today, AI accelerators represent mid-teens of TSMC’s total revenue.

Another interesting data point reflecting the AI growth for TSMC is its customer base, and for the first time in 10 years, Apple isn’t its largest customer and was overtaken by Nvidia, which alone represented 41% of TSMC’s incremental revenue growth.

I wouldn’t be surprised for the gap between Nvidia and Apple to widen for years to come, and internally, I expect Nvidia to reach 25% of TSMC’s revenue by the end of the decade, even as it diversifies its own fab supply chain.

Beyond Nvidia, the industry teardown of other major chips points towards more growth for TSMC. In custom silicon, Amazon’s highly successful Trainium 2 chips are rumoured to leverage TSMC’s CoWoS advanced packaging technology, while TSMC also fabricates Google’s 8th generation of its Tensor Processing Units (TPUs). Altogether, this brings us to even more growth for 2026 and beyond, bringing us to TSMC’s updated valuation.

TSMC’s updated valuation model

Read more

Register for call

Later this week, on Thursday, I’ll be hosting the Q1 2026 Investor and subscriber virtual call.

The investor call is an opportunity to hear directly from me on recent portfolio and investment updates. In Q1 2026, we published deep dives on Visa, Mastercard and Microsoft.

There will also be a Q&A, and you can either send your questions via the comment section below, in the registration form attached below or during the call.

To register for the virtual call, please fill out the registration form on the Jenga IP website here.

Date: 30th April 2026

Time: 17:00 - 18:00 (GMT)

Location: Virtual

Last week, I published a 22,000-word deep dive on Microsoft. Here’s the one-page version.

Company overview

• Founded in 1975, Microsoft is the largest enterprise software company, operating in the office suites and productivity apps, ERP, CRM and database software markets.

• It’s also the second-largest cloud services provider and continues to diversify its solutions across both traditional and AI cloud services.

Business quality and track record

Read more

While digging into Microsoft over the past month, some insights particularly stood out, and one of these was the staying power of software once they become embedded into workflows. There’s a group of five listed companies founded between 1972 and 1983 that best exemplify this: Microsoft, SAP, Oracle, Adobe and Intuit, what I call the foundational five software companies.

To highlight their dominance, the tables below compare the foundational five companies with the 773 listed IT software companies (GICS).

As highlighted in the table, the foundational five companies achieve EBIT margins (32.8%) 6 times greater than the listed IT software industry and surprisingly, they also achieved faster revenue and operating profits growth over the past seven years, with their 7-year EBIT growth being twice as fast as the industry, despite their size.

The second table contains one datapoint I want to focus on a bit, their share of the industry’s net income, currently at 90.2%. This highlights how oligopolistic and winner-take-most software is. While some might argue this is normal for a mature industry, no other industry comes close to this level of concentration from a public market’s perspective.

The table above compares the share of net income the largest five companies represent, and as you see, the next highest concentration is pharmaceuticals at 43.9% share, still less than half of software’s 90.1% share. Software may be the most winner-take-most industry and the real question today is to what extent and if AI breaks this structure.

A reasonable pushback is that a significant portion of Microsoft’s profits comes from cloud services, which arguably sit outside traditional software. Even excluding Microsoft entirely from the dataset, the remaining four companies still account for 74% of IT software net income, well above any other industry.

Adobe and Intuit

Until recently, the five companies seemed indestructible, with both Adobe and Intuit’s forward P/E exceeding 50x in 2021.

The AI challenge has brought a needed reality check to their valuations; Adobe, for example, declined to a 20-year low forward P/E of 9.8x while Intuit, for the first time since 2009, trades at a forward P/E discount to the S&P 500 Index.

Over the next few weeks, I’ll be spending some time researching both companies in-depth and will publish deep dives, starting with Intuit.

At the moment, I only own shares in Microsoft among the five companies and given how exposed the portfolio is to potentially at risk from AI companies (Microsoft, Visa, Mastercard and FactSet - 30% weighting), I’m raising my investment IRR hurdle from 15% to 25% for new software investments.

Beyond these companies, there are a few other companies on my watchlist, due for further research. During the Microsoft deep dive, I created a top-down overview of software, splitting the industry into 42 sub-industries and ranking them on key attributes like barriers to entry and market concentration.

Some of the watchlisted to-be-deep-dived-if-the-price-becomes-attractive companies include ServiceNow, Fortinet and Salesforce.

Finally, I’m also working my way through all 773 listed IT software companies from A-Australia to U-United States. If any other interesting companies that I understand, can value and seem to have favourable risk/reward come up, you can expect to receive additional research on them.

Due to the length of the report (75 pages), a more accessible PDF version is attached at the end of the article, alongside the software industry, data centres and cloud services data sheets.

Like many readers and value investors, I’ve been carefully watching the current software downturn, researching and pondering on how these companies might evolve in a more AI-driven world.

Admittedly, I’ve become more wary about the future of software. Until recently, software seemed like a no-brainer. These businesses were asset-light, highly profitable, with low incremental costs and could grow via both organic and M&A channels.

Although far from being mutually exclusive, AI challenges the traditional economics of software by directly targeting both the number of employees at companies (volume, fewer seats) and revenue charged per user (price, lower ARPU), driven by the potential productivity gained from AI across teams, the unbundling of previous distribution channels and increased competition for a share of technology budgets.

Of course, not all software is created equal, and AI will further enable some. Factors such as workflow embeddedness, providing mission-critical solutions, having solution offering breadth, being financially robust and disciplined will differentiate software companies. After further research and assessing an initial shortlist of software companies, I concluded that the largest, Microsoft, had an attractive risk/reward and initiated an investment position in its shares.

Market capitalisation (As of 9th March 2026): $3.04 trillion

Jenga IP 2030 FY estimated market cap: $5.6 trillion

Potential IRR (including dividends and buybacks): 16.7%

Jenga IP Quality Rating: 84.2/100

More specifically, the Microsoft investment case rests on five key questions:

The Microsoft investment case questions

• Can Microsoft maintain its software margins despite a more capital-intensive business model?

• Will Azure remain a high-margin business or evolve to utility-like infrastructure economics?

• Will Copilot meaningfully expand the ARPU potential across Office 365, or does customer seat loss prove more meaningful to its bottom line?

• Can Microsoft retain control of enterprise workflows, or does AI shift value to new platforms?

• Does AI strengthen or weaken Microsoft’s moat due to software and workflow commoditisation?

To answer these five questions, here’s what I’ll cover in the Microsoft deep dive:

Table of contents

• History of Microsoft: A summary of Microsoft’s history, efforts in building enterprise software across its four management phases (Bill Gates, Steve Ballmer, Satya Nadella 1.0 and 2.0). Case studies on Microsoft’s challenges over the years and insights on its transition from the license model to SaaS, its early AI efforts, and a conclusion on the current state of its competitive advantages (distribution and bundling, balance sheet and Azure’s switching costs).

2. Overview of Microsoft’s services: A breakdown of Microsoft’s product and service offerings and their performance relative to competitors, further overview of the past transition from license to SaaS and benefits gained, a deep dive into Azure’s cloud offerings and market position, Copilot and Microsoft’s broader AI offerings in focus and the enterprise perspective of Microsoft (M365, Azure, Dynamics, security, Windows and GitHub & developer tools).

3. Enterprise software industry: An overview of the global enterprise software industry and a deeper analysis of Microsoft relative to other first-generation software companies (Oracle, SAP, Intuit and IBM), the broader commercial software business and productivity tools market (Alphabet Workspace and WPS) and the challenge from AI-native companies (Anthropic and OpenAI) to traditional enterprise software.

4. Inside Microsoft’s AI and data centre efforts: Insights into the current data centre and cloud computing market, a value chain overview from land permit to deployment, assessment of current competitors (Amazon, Alphabet and neocloud players), token economics and the Microsoft Fairwater strategy, the developer platform and its role in Microsoft’s AI efforts (GitHub Copilot, Visual Studio and OpenAI partnership) and the shift in market dynamics by AI layer (infrastructure, platform, applications and control points).

5. Business economics: A deep dive into Microsoft’s business economics by revenue (service type, business segment and geography), costs (cost of revenue, operating costs and Capex), assessment of the current state of Microsoft 365 competitiveness and strategy, Azure and Azure AI’s efforts and unit economics and a competitive landscape and further insights into LinkedIn and Gaming.

6. Growth opportunity: Analysis of its growth drivers (seat, ARPU and usage) across each business segment (server products and cloud services, commercial - insights into E5 to E7 and Copilot - and consumer products and cloud services, LinkedIn, Gaming, Windows and Devices, Search and News Advertising, Dynamics and Enterprise services), an overview of the changing cost profile to its operating margins.

7. Risks and challenges: Further assessment of Microsoft’s transition into an asset-heavy business model, coopetition with OpenAI and Nvidia, an overview of the talent market among Big Tech companies, AI operational gaps to Alphabet, impact of a potential slowdown in AI at the enterprise level and concluding thoughts on its evolving value chain and other risks like cloud commoditisation, pricing pressure, lack of frontier model ownership, changing software architecture and the evolving regulatory landscape.

8. Valuation: An earnings multiples-led valuation and projection of Microsoft’s earnings potential through to 2030 FY, comparison with peers and the broader software industry.

9. Conclusion: Final thoughts on the broader software market downturn and concluding answers to the initial five Microsoft questions.

Subscribe now

1. History of Microsoft

Microsoft’s journey to becoming among the greatest businesses of all time has been well documented in books, media and documentaries. Rather than attempting another history lesson, to stay relevant to its investment case, it’s more useful to focus on the competition and challenges it’s faced to date. To achieve this, I will assess Microsoft through the eyes of its competitors and partners over the past three leadership reigns: Bill Gates, Steve Ballmer and Satya Nadella.

Bill Gates (1975 - 2000)

IBM (1980)

Although Bill Gates and Paul Allen had founded Microsoft five years before the key 1980 partnership with IBM for IBM’s PCs, the contract was the pivotal start for Microsoft in the operating software market. As described by the media, IBM was the “sun, moon and stars” in providing technology solutions to businesses and to defend itself against growing competitors, IBM launched “Project Chess” to build a PC suitable for businesses, and after failed negotiations with CP/M, IBM turned to Microsoft to build its operating software. Microsoft didn’t actually have an operating system, and Bill Gates swiftly acquired all rights to 86-DOS from Seattle Computer Products and then repackage the OS for IBM.

Beyond being technically savvy, Bill Gates was also a capital allocator and had a strong sense of acquisitions that could bolster Microsoft’s software position, and when both skills are combined, the opportunity is limitless. The IBM partnership propelled Microsoft, and between 1979 and 1985, Microsoft’s revenue grew at a 97% CAGR to $140.4 million with an impressive EBIT margin of 29%.

IBM had also benefitted from the partnership and sold over 500,000 PCs over the first 2 years, becoming the market leader, but the mistake they made was that they underestimated Microsoft’s ambitions, the value of owning your operating software and the commodity nature of PCs. As Microsoft began expanding its own commercially successful Windows OS during the mid-1980s, tensions increased between IBM and Microsoft and by 1990, IBM ended its relationship with Microsoft, deciding to take full control of the OS/2 and build its own OS internally.

IBM faced a more important challenge. Its PC architecture was too easy to clone, leading to many new OEM entrants into the PC market. By the end of the 2000s decade, Microsoft had eclipsed IBM’s operating profits, from being 4% of its EBIT in 1990.

“Eventually we need to have at least a neutral relationship with IBM.” Our strategy is Windows, 1991 memo by Bill Gates

OEMs (1982)

In the 10 years to 1985, PCs in America grew at an astonishing 98% growth rate, and while IBM led the market, it didn’t grab the profits or majority share because its BIOS firmware which operated between the hardware and software was easy to copy and new entrants like Compaq, Dell and HP cloned IBM’s structure, leading to a fairly commodity PC market, pricing wars and low profit margins. As we have previously learned, growth alone isn’t enough.

Microsoft made a crucial business decision with OEMs and made its OS licensable for a fee. This was a once-in-a-generation period for OS because Microsoft had very little competition in the OS market; Apple had already decided to operate a closed ecosystem PC model, and the competition intensity in the PC market meant it was far more economical for these new OEMs to pay Microsoft than to attempt developing their software in-house.

Although selling OS through the software channel had a lower margin when compared to direct to consumer, the former had higher volume, required less marketing costs, which was in line with Bill Gates’ philosophy of “volume, not unit price.”

Unlike the PC market, the OS market had more of a winner-take-most economics, and the intense competition tilted the industry’s value chain towards Microsoft. To date, the OS-PC relationship has remained in Microsoft’s favour, and its biggest competitor wasn’t another OS company, but rather, a PC manufacturer that continued to develop its own OS, Apple.

Apple (1984)

Before 1984, Apple had already licensed Microsoft BASIC for the Apple II, but the real gains for Microsoft began with Microsoft Office, particularly the Excel spreadsheet application. Before Excel, Microsoft had developed Multiplan to compete in the spreadsheet market with the market leader, Lotus 1-2-3’s (later acquired by IBM in 1995). Struggles with the interface and functionality impacted the commercial success of Multiplan.

An extension of its relationship with Apple (first with Microsoft Word) brought Microsoft closer to what the future graphical interface could look like, and Microsoft began extending its applications for Apple’s Macintosh. Excel proved successful and was a hit with Macintosh users, with Microsoft flagging this as the primary driver for its growth in 1985 sales in its 1986 IPO prospectus.

“Operating results for 1985 were influenced by the introduction of new products for the Apple Macintosh (Microsoft Word, Microsoft File and Microsoft Business Pack.” - Microsoft’s 1986 IPO Prospectus.

Similar to IBM, Apple underestimated Microsoft’s ambitions, and Microsoft later developed a new operating system that claimed to have stolen several features from Apple’s Macintosh OS, leading to a legal battle by the end of the 1980s. Apple would later lose the case to Microsoft, face dwindling sales, get saved by Microsoft with a $150 million equity investment and then later surpass Microsoft in market cap, revenue and net profits, remaining a key competitive threat to Microsoft.

“Microsoft versus Apple has been such a prominent and even contentious rivalry that people forget we’ve been building software for the Mac since 1982.” Satya Nadella. Hit Refresh: A memoir by Microsoft’s CEO.

Wars: Oracle (1989), Intuit (1992) and Netscape & AOL (1995)

Create, bundle and compete

By 1990, Microsoft had just surpassed the $1 billion revenue mark, but the growing competitive threats and innovations, particularly in application software, further expanded PC use cases beyond enterprise, and most importantly, the internet, meant Microsoft’s software ecosystem control was under threat. To ensure it remained ever relevant to customers, Microsoft further doubled down on its product suite and solutions.

While remaining focused on software, it diversified by expanding its application software presence via acquisitions like Forethought (1987), which became PowerPoint, launched Internet Explorer (1995) to compete with Netscape in the browser market, hotmail.com (1997) and Navision (2002) which evolved into Microsoft Dynamics, entering the database market through a partnership with Sybase in 1989, competing with Oracle and launching new products like Microsoft Money which attempted to compete with Intuit.

Microsoft had some spectacular failures, such as Microsoft Money, which resulted in a 1994 attempt to acquire Intuit for $1.5 billion. That said, there were also wins where Microsoft moved very aggressively, such as the browser war with the fast-emerging Netscape Navigator. Despite the fast internet growth, it was still unclear how companies would eventually monetise the internet, and the browser being the first point of entry to the internet seemed attractive, and to ensure victory here, Microsoft aggressively bundled the Internet Explorer with the very successful Windows 95 OS, free of charge to customers.

AOL would later acquire Netscape, but the aftermath of Microsoft’s aggressive business tactics led to the dark antitrust years, a period Bill Gates claimed distracted Microsoft and cost it the loss of the mobile revolution.

Steve Ballmer (2000 - 2014)

The Steve Ballmer years as CEO of Microsoft were often summed up as the dark Microsoft days, a 14-year period its share price declined by 33% and experienced costly failures such as missing out to both Apple and Google in mobile computing, missing out on the advertising-led internet business models to Meta Platforms (Microsoft attempted to buy Facebook) and Google and internal product failures with Windows Vista and the most financially costly mistake, its efforts with search via Bing.

During these 14 years, Microsoft grew its net profits by 6.3% annually, and while this might seem like a total failure, Microsoft began building the foundations for a more enterprise-centric company.

Oracle (2000)

Winning the enterprise customer

Contrary to popular thought, Microsoft wasn’t initially a dominant force in enterprise solutions as it is known today. In the early 1990s, to extend its solutions beyond operating software, Microsoft entered the relational database market, which helped businesses store and retrieve structured data more easily via a range of applications like Microsoft Access, Microsoft SQL and Microsoft FoxPro.

Oracle had long dominated this market after creating the first relational DBMS in 1979, but as more players began entering the market, they diversified into enterprise software, competing more directly with Microsoft. The bitter rivalry between the two companies would take a turn when Oracle’s founder, Larry Ellison, admitted to hiring private detectives to spy on Microsoft and Bill Gates.

More relevant to the investment case today is while brutal at times; in 2003, both companies competed directly for 94 contracts, Microsoft’s server operations and competition with Oracle pushed them into being more enterprise-oriented; understanding client needs, enterprise sales and tender processes, technology consulting (Microsoft founded a successful consulting JV with Accenture in 2000 - Avanade), improving its brand and corporate image and trust with major corporations, among others.

Steve Ballmer’s push in the server and tools market proved quite successful, and by 2007, it represented 22% of Microsoft’s revenue. More importantly, it prepared Microsoft for the next phase of computing, selling the cloud to enterprises.

Amazon Web Service (2008)

Willingness to self-disrupt

Today, Microsoft’s cloud computing division represents 38% of the group’s revenue, but years ago, it was a startup project within its server and tools division. Several technology companies had already begun investing in cloud solutions, particularly Amazon with its AWS Elastic Cloud solutions. The challenge many older software companies faced, like Microsoft, Oracle and VMware (later acquired by Broadcom), was that cloud computing cannibalised their existing operations, and Microsoft was among the few willing to self-disrupt quickly.

The Azure cloud platform saved customers the need to build their own data centres, and alternatively, rent compute and storage bandwidth from Microsoft on a pay-as-you-go basis. As we’ll discuss later, Amazon was more aggressive with scaling its cloud platform, particularly with startups and small businesses and has maintained its market-leading position in the cloud market since its inception. That said, Microsoft has stayed a durable number two, strategically leveraging its trust with larger corporations, including corporations like Boeing, Pixar and Toyota as initial customers.

Although the economics of cloud weren’t as attractive as Windows OS or Office, it kept Microsoft at the forefront of technology and also supported the broader transition of the rest of Microsoft’s businesses from the license model to the cloud subscription format.

Although several people were critical of Azure’s growth (code-named Red Dog), like Ray Ozzie, Mark Russinovich and Dave Cutler, Satya Nadella, then president of the broader Server and Tools division, was selected by Ballmer to lead Microsoft as its CEO.

Satya Nadella 1.0 (2014 - 2022)

Microsoft was no longer the only golden technology company and had to share technology leadership with new players like Alphabet and Meta Platforms, and to prepare for this era, Nadella made conscious efforts to broaden the corporate culture within the server and tools division, a market where it wasn’t the incumbent market leader, by focusing on win-win partnerships. Microsoft’s business relationships with rivals like Adobe, Samsung and Alphabet were rebuilt, avoiding costly litigation and also supporting industry peers building products for its broadened ecosystem.

Another key factor central to the investment case for Microsoft is customer stickiness and the transition from the license model to subscription, which, in hindsight, proved to be extremely beneficial for the group.

Case study: The license to the SaaS model transition

One of the ten categories of quality is the nature of demand, companies with products and services that are mission-critical, profit margin resilience during downturns, and recurring solutions. Microsoft had always been mission-critical, but the perpetual license of most of its services meant it lacked the other attributes, which made sales lumpier despite a minimal marginal cost to serve customers.

By 2008, Microsoft had already started exploring the subscription transition through Microsoft Exchange and Microsoft SharePoint, and by 2010, it launched Office 365, bringing the rest of its productivity tools to customers on a subscription basis, initially priced at $6 per user per month, in line with the Office pricing, assuming a 3- to 5-year replacement period.

On the cost side, the shift to SaaS required more investments in its data centres, compute and other delivery costs such as monitoring, security and marketing to reduce churn rates, leading to an initial profitability burden. As an organisation, Microsoft’s EBIT margins continuously declined from 39.8% in 2008 to 29.8% in 2016 (chart below), with operating profits growing by just 1.5% annually. Of course, the heavy investments required in delivering the Azure platform alongside hardware bets like the Nokia acquisition and the Microsoft Surface laptops partially contributed to the margin dip during this period.

As time would prove, the transition to the SaaS model (operationally and culturally) proved highly beneficial and accretive to its prospects, particularly when compared to its 1970s peers like Oracle and SAP, given its faster and more decisive execution.

AWS competition (2014)

Letting go of some dogmas

Given the eventual success with the Azure cloud platform, one might think Microsoft always got its strategy right from day one, but they were actually forced to evolve from a platform-as-a-service model (PaaS) to a broadened Infrastructure-as-a-service (IaaS) model.

Microsoft had previously viewed open-source as a threat to Windows, with Ballmer going as far as calling Linux, the dominant operating system for cloud servers, a cancer. However, to dominate cloud economics, Microsoft needed to embrace open-source technologies, an approach AWS had fully embraced, supporting its client wins. By 2012, Microsoft began reversing past decisions, launching Linux Virtual Machines (VMs), and within five years, 40% of all VMs were running on Linux.

Previous foes like Oracle, SAP and Red Hat Enterprise also extended their tools and services to Azure, making Azure a full IaaS competitor to AWS. Other integrations like Xamarin and acquisitions like GitHub (2018), LinkedIn (2016), which forced Microsoft to open APIs at a massive scale, transitioned Microsoft to a more open-platform business model, preparing the group for the technology transition, positioning it for AI.

Satya Nadella 2.0 (2023 - present)

Artificial intelligence

Before OpenAI’s ChatGPT launch in November 2022, Satya Nadella had long voiced his optimism for AI, as early as 2015, but before this, Microsoft’s AI efforts and financial results were limited to Azure and gaming experiences like its Halo franchise. Several failed products like Cortana (voice assistant) and Zo (chatbot) showed Microsoft wasn’t a serious player in AI, but that changed in 2023 when Nadella reframed Microsoft as an AI-first company.

Today, the narrative on Microsoft’s valuation is its AI competitiveness relative to big tech peers and startups and given recent share price drawdowns, it’s clear that the market thinks negatively of Microsoft’s position. I’ll discuss this more broadly in this deep dive, but first, it’s important to mention Microsoft’s efforts so far:

• Integrating AI to product portfolio: The highest ROI and accretive step in most technological transitions is adapting existing product suites for AI, and Microsoft has moved both too slow and too fast to the extent it’s had to remove AI integrations (Copilot) from its Office and Windows 11 products. After several iterations, it seems Microsoft has now consolidated all its efforts under the Copilot franchise, serving as a companion to existing users. Unlike Google, Microsoft hasn’t established its own LLM (Large Language Model) and, rather, has focused on SLMs and integrating LLMs from peers into its Copilot offerings. I will discuss the challenges of this strategy later on.

• Azure: Azure and cloud computing services is natural beneficiary of AI because they increase demand for compute, storage and networking. Beyond the Azure infrastructure, Microsoft has added more models for its enterprise customers, integrated apps and agents (Microsoft Foundry, GitHub) for customers, and, when combined with its governance and compliance standards, it has narrowed the cloud gap to Amazon.

• External investments: Externally, Microsoft has invested in other companies, with OpenAI being the most significant, now valued in excess of $100 billion. Microsoft itself was built on these partnerships (IBM), and to ensure its further tied to AI’s growth, it has extended beyond just service provision to equity investments. There are risks worth discussing here, and I will highlight them in sections 4 (Inside the Data Centre) and 7 (Risks and Challenges).

Beyond the cloud, Microsoft has less room for product monetisation from AI when compared to Meta, Alphabet, and Apple, and integrations and partnerships will be even more critical here. A key pillar of Nadella’s tenure has been repairing Microsoft’s image to the broader ecosystem, and I believe this will further support its growth opportunity.

Current state of Microsoft’s moat

In the wake of another technological shift (AI everything), it’s tempting to doubt once again Microsoft’s ability to remain relevant, and here, I believe mapping out the historical context as we have just done, studying past failures, lessons learned and assessing the results of past decisions is key to gaining clues on how the future might look. Some key competitive advantages for Microsoft in this shift include:

• Distribution and bundling power: Although Microsoft’s AI strategy still seems somewhat scattered (it partially is), distribution and bundling, two competitive advantages that have become even more effective today than ten years ago, are key competitive advantages for Microsoft. We got to see how effective Microsoft’s bundling of Teams, its web conferencing app, with the rest of the Office offerings was in neutralising Zoom, its main competitor’s rise. There’s still lots of scope for further product bundling when the enterprise customer is taken into consideration, and I believe its existing distribution to 400 million+ users will prove valuable.

• Azure’s switching costs: The transition to SaaS in the enterprise software market has effectively reduced its once impenetrable switching costs, but in the cloud, AI further complicates the challenges, time, costs and resources required with switching cloud providers or taking certain processes in-house. Although cloud computing is mainly on a pay-as-you-go basis, the switching costs in time and effort remain immense.

• Balance sheet strength: During the current AI boom phase of excessive funding, having balance sheet strength might seem irrelevant but as we saw during the 2000s tech bubble, Apple and Microsoft’s survival versus the hundreds of startups like Netscape, balance sheet strength will always be a key source of competitive advantage and when coupled with capital allocators willing to combine technological optimism (Satya Nadella - CEO) with discipline (Amy Hood - CFO), I believe it shapes the long-term investment case.

The table below highlights some Microsoft failures over the past decades. Despite these 13 big failures that cost them hundreds of billions in lost revenue, it’s impressive that Microsoft still manages to be a $300 billion+ revenue company, which highlights a final key point: its ultra-wide addressable market that keeps the growth potential huge.

Subscribe now

2. Overview of Microsoft’s services

With Microsoft’s historical context now set, I will next introduce the products and services that make up Microsoft, discussing some learnings from my review of its various products. At a high level, I divide Microsoft into two parts: intelligent cloud (35% of revenue) and non-intelligent cloud (65% of revenue).

Microsoft’s non-intelligent cloud offerings (65% of revenue)

We can refer to this as the “traditional and acquisitions” segment of Microsoft. In today’s AI-driven world, the economics and growth prospects of this segment have taken a back seat among market commentators despite representing the larger share of Microsoft’s revenue and 65% of its operating profits. The non-intelligent cloud is split into three segments:

• Enterprise Applications (Microsoft 365, Power BI, Dynamics 365)

• OS and devices (Windows OS and Surface devices)

• Consumer platforms (Gaming, Search and News Advertising and LinkedIn)

The table above maps out the core applications within each of these segments, and next, I’ll introduce the key products across these segments, starting with the largest, Microsoft 365 (commercial and consumer).

Microsoft 365

Microsoft’s 90%+ market share in operating software often took the limelight during the 1990s, but a lesser-known fact is that application software mattered just as much to Microsoft over the years, and while historically less profitable than operating software, it was less cyclical and more resilient to economic and technology industry shocks.

As the transition to the SaaS model opened a new growth engine with innovation and product bundling, the applications division has become significantly more important than OS. I estimate it is now nearly 6 times larger in revenue than OS, from roughly the same size during the 1990s and early 2000s. This is an important lesson on the benefits of diversification in technology and a platform-like business model.

Let’s next look at the three important application segments:

• Core apps (Word, PowerPoint and Excel)

• Backbone and control layer (Outlook, Exchange, SharePoint and OneDrive)

• Interaction layer (Teams)

Core apps (Word, Excel and PowerPoint)

The three applications are central to Microsoft 365 and are utilised by over 500 million customers annually with Word (documents), Excel (spreadsheets), PowerPoint (presentations), each addressing the key workflow and productivity tools and while not officially provided by Microsoft, I estimate they contribute directly around half of Microsoft 365’s total revenue, or $42 billion annually, averaging an EBIT margin of around 73% (Jenga IP estimate).

Microsoft's Product Portfolio

In the 1980s, Microsoft realised the potential value each of these markets could become and given the nature of its “winner-take-most” economics they responded quickly with creating the three applications over time, introduced file formats such as .doc for Word and .xls for Excel, increasing the switching costs to rival products built by Lotus 1-2-3 and then bundled each into the Microsoft Office product.

The Office 365 consumer plans are currently priced between $4.4 to $18.8 per user per month, while one-time Office 2024 currently retails for $180. When scored across 13 product metrics (recurring demand, Big Tech competition, benefits from bundling, market share), the three core apps score very favourably compared to the other 365 applications. You can see this from my colour-coded (0-10) table below.

While all three continue to dominate their productivity categories (see table below) and are used across all industries and customer groups, there’s some competition, led by Google’s Workspace, who have won customer end markets like startups and the education end-markets.

Read more

Visa is a high margin and capital light payments network with durable double-digit earnings growth, pricing power and significant capital return with a 16.4% IRR potential over the next five years.

Company overview

• Founded in 1958, Visa is the leading global payment network infrastructure that provides issuers, merchants, acquirers and consumers with a central network to transact and process payments for goods and services.

• Its network authorises, settles and clears payment transactions, and for solutions provided, Visa charges a network/scheme fee (approximately 0.5% of the average transaction value).

Read more

I had planned to complete the fourth deep dive into non-banking financials with a primer into the insurance industry, alongside a follow-up deep dive on Fairfax Financial Holding (we already looked at Farmer Mac, FactSet and Visa/Mastercard)

After this, I had also planned to publish a premier on the credit reporting industry with a follow-on deep dive covering Experian, and then tackle our first bank investment research, NuBank and the broader Brazilian banking sector.

Developments in the software industry have, however, meant I’ll need to shift my attention, given the recent valuation drawdowns seen across the industry.

Even though I’ve published so far successful investment deep dives on other technology companies like Alphabet (March, 2025 - now exited) and TSMC (July, 2025 - ongoing investment position), I certainly don’t consider myself a technology or software expert but a generalist searching for value in quality companies.

As we identified during the global industry overview, software has produced some of the most durable, high-performing and profitable companies of all time, 2nd only to the tobacco and beverages sub-industry among all 60 sub-industries globally (GICS).

A handful of software companies have surprisingly found ways to survive several technology shifts. Since being born out of IBM, SAP continues to lead the Enterprise Resource Planning (ERP) market, 54 years later, with many other software sub-markets still dominated by the class of 1970s/early 80s companies (Oracle, Intuit, Apple, Adobe, Bloomberg and Microsoft).

I’ve learned from many great investors never to waste a drawdown, and I will be spending the next 150 hours researching Microsoft.

I’m currently finding similarities (and some differences) between the 1970s IBM-Microsoft relationship and the current Microsoft-OpenAI relationship, and the mainframe-to-PC transition, with the client-server to GenAI transition, and as I found during the Alphabet deep dive, I believe there are many learnings from the past that we can apply to how one assesses big tech today.

I also plan to publish another software deep dive after Microsoft (company yet to be decided, but Adobe is a challenger - no position), and premieres on both AI data centres and challenges for enterprise software in an LLM world.

For premium monthly readers who signed up to follow our financials research, I suggest pausing your subscription until May, when I plan to continue our financials research. For more generalist readers, we are about to get into the weeds in software. Here’s what I’ll cover in the Microsoft deep dive:

Subscribe now

Microsoft’s deep dive table of content

1. History of Microsoft: A summary of Microsoft’s history, efforts in building enterprise software and beyond, across its four phases (Bill Gates, Steve Ballmer, Satya Nadella 1.0 and 2.0) and lessons from IBM’s struggles. Case studies on Microsoft’s challenges over the years and insights on its transition from the license model to SaaS and its early AI efforts, and a conclusion on the current state of its competitive advantages (distribution, bundling, balance sheet and switching costs).

2. Overview of Microsoft’s services: A breakdown of Microsoft’s product and service offerings and their performance relative to competitors, further overview of the past transition from license to SaaS and benefits gained, a deep dive into Azure’s cloud offerings and position, Copilot and Microsoft’s broader AI offerings in focus and the enterprise perspective of Microsoft (M365, Azure, Dynamics, security, windows and GitHub & developer tools).

3. Enterprise software industry: An overview of the broader enterprise software industry and a deeper analysis of Microsoft relative to other first-generation software companies (Oracle, SAP, Intuit and IBM), the broader commercial software business and productivity tools market (Alphabet Workspace and WPS) and the challenge from AI-native companies (Anthropic and OpenAI) to traditional enterprise software.

4. Inside Microsoft’s AI and data centre efforts: Insights into the current data centre and cloud computing market, a value chain overview from land permit to deployment, assessment of current competitors (Amazon, Alphabet and neocloud players), token economics and the Microsoft Fairwater strategy, the developer platform and its role in Microsoft’s AI efforts (GitHub Copilot, Visual Studio and OpenAI partnership) and the shift in market dynamics by AI layer (infrastructure, platform, applications and control points).

5. Business economics: A deep dive into Microsoft’s business economics by revenue (service type, business segment and geography), costs (cost of revenue, operating costs and Capex), assessment of the current state of Microsoft 365 competitiveness and strategy, Azure and Azure AI’s efforts and unit economics and a competitive landscape and further insights into LinkedIn and Gaming, and an overview of recent hyperscaler Capex growth trends, and its risks.

6. Growth opportunity: Analysis of its growth drivers (seat, ARPU and usage) across each business segment (server products and cloud services, commercial - insights into E5 to E7 and Copilot - and consumer products and cloud services, LinkedIn, gaming, Windows and devices, search and news advertising, dynamics products and enterprise services), an overview of the changing cost profile to its operating margins. Case studies on past capital allocation, M&A optionality, and distribution as a moat.

7. Risks and challenges: Further assessment of Microsoft’s transition into an asset-heavy business model, coopetition with OpenAI and Nvidia, an overview of the talent market among big tech companies, AI operational gaps to Alphabet, impact of a potential slowdown in AI at the enterprise level and concluding thoughts on its evolving value chain and other risks like cloud commoditisation, pricing pressure, lack of frontier model ownership, changing software architecture and the evolving regulatory landscape.

8. Valuation: An earnings multiples-led valuation and projection of Microsoft’s earnings potential through to 2030 FY, comparison with peers and the broader software industry.

9. Conclusion: Final thoughts on the broader software market downturn and concluding answers to the initial five Microsoft questions.

During my weekly Saturday morning checkup on portfolio companies, I came across a Reuters article that mentioned two private equity firms, Thomas Bravo and Hellman & Friedman, are currently looking at FactSet as a take-private acquisition.

At the time of my research into FactSet, while I considered the potential of a takeover given its mid-cap size, the lack of a founder or controlled shareholder, its business quality and current valuation, my investment thesis was solely a long-term buy and hold case.

I’m not a fan of speculative analysis but given Reuters is a very credible source, the article mentioned “three people familiar with the matter”, Thoma Bravo recently raised $34.4 billion for its Fund XVI and Discover Fund V and has already displayed its “deal hunger” with a recent acquisition of WWEX and a $12.3 billion acquisition of the previously listed HCM software company, Dayforce.

Hellman & Friedman, on the other hand, seems less deal active at the moment, but after a successful investment in Allfunds, now in the process of a takeover by a strategic investor, Deutsche Börse, one can also assume they’re also looking at new investments in the sector.

Admittedly, my preference is for these to be simply rumours. A 25% premium from FactSet’s current share price would actually be a failure for us, given the opportunity cost in both capital and time, so as you can imagine, I’d prefer for FactSet to stay public.

That said, there’s little we could do with less than 0.005% ownership of FactSet, so it’s important to consider this possibility, and the rest of the article will assess a potential takeover more broadly from three lenses: management, the potential buyer and current shareholders.

1. FactSet’s management view

FactSet’s management has two options:

• Actively seek the best buyer

• Strongly refuse to sell

In my view, the key defining factor that should determine FactSet’s management decision should be its confidence in growing net profits and free cash flow above 6-8% annually for the next 4-5 years.

I would suggest that if management is highly confident in its ability to grow FactSet by at least 6-8% over the next few years, primarily through subscription growth, staying public would be more accretive to both management and shareholders.

However, if management has low confidence in its ability to deliver net profit growth by at least 6% in the medium term, selling FactSet now is the better decision (I hope this isn’t the case).

FactSet’s issue today isn’t profitability; it is earnings growth and resilience, and proving its strength here, in my view, is the single most important factor between a 13x P/E and 21-23x P/E valuation by the market.

2. The Private equity’s view

I’m not surprised by the interest from private equity firms:

• FactSet retains 90% of its clients annually, with 95% retention for its annual subscription value.

• It is mission-critical to its customers.

• It’s highly profitable, cash generative and maintains a good balance sheet with further leverage potential, as is typical with PE.

• Private equity is a major client of FactSet, and as owners, they would likely share creative operational ideas on how best to deliver earnings resilience. I suspect Preqin and BlackRock are an ongoing case study of this with the recent BlackRock Aladdin integration.

• The purchase price is low compared to alternatives. At a 25% premium to today’s market cap, FactSet would be valued at. For perspective, although a totally different business model and financial profile, Thoma Bravo’s listed SailPoint is currently valued at 27x its 2028 market estimated EV/EBITDA.

• Given evidence of both private equity and financial exchanges’ long-term interest in financial market data companies, should a PE owner prove successful during a 6 to 8 year ownership period, there would be a few exit options available, including a re-IPO, with IHS Markit serving as a good case study.

• Finally, Thomas Bravo’s portfolio company, BlueMatrix, recently acquired RMS Partners from FactSet, and one can assume there’s already a relationship between both companies from this deal. Achim Fehrenbacher, a former 20-year FactSet veteran now works with Thoma Bravo’s BlueMatrix, and I suspect Scott Crabill, a Thoma Bravo managing partner, would be leading any deal with FactSet. After assessing Thoma Bravo’s 17 current Application Flagship investments, I don’t see any other similar financial market data investments and given their track record of merging similar companies, this could be a downside to any take-private transactions.

Looking more broadly at the listed U.S. financial information services space, the PE firms currently have 3 take-private potential targets:

• Morningstar: Among the three companies, Morningstar has the most resilient business model given its diversified products and client base, growth potential, and its balance sheet is currently PE-friendly. The challenge they’ll face is the founder and chairman, Joe Mansueto, a 45% owner of Morningstar. Having watched some of his interviews and read some shareholder letters from its CEO, I doubt he’s likely to sell. A PE buyer would also have to purchase Morningstar at a more premium multiple and at a higher enterprise value (>$10 billion).

• Gartner: Compared to the other two companies, Gartner’s services are most likely the least sticky (85% client retention rate), have the slowest organic growth potential, the most leveraged, and also most at risk to the ongoing AI challenges. Like FactSet, though, Gartner isn’t founder-owned, with its long-time CEO, Eugene Hall, owning roughly 1.7% of its shares, which eases a deal process.

• FactSet: FactSet isn’t as high quality and diversified as Morningstar and potentially faces more AI threats with a slower growth potential, has limited room for margin expansion but that said, its lower purchase price and more simplified acquisition process likely make it the most attractive of all three companies from a PE perspective.

As a result of this, I think the possibility of interest from private equity firms is quite real, and if I were to speculate on reasons why it was leaked to Reuters, the most interested buyer (I suspect this is Thoma Bravo) is probably looking for another PE firm to partner with for the deal, given the enterprise value.

Given that Thoma Bravo is also the largest software PE firm and several rivals like Apollo and Blackstone have taken a stance against software companies, I suspect there’s limited competition in a FactSet acquisition.

Of course, these are just my thoughts; I have no insider insights here.

3. The shareholder’s view

FactSet has a very passive shareholder base, with the majority of its shares held by big American passive institutional funds. BlackRock, Vanguard and State Street, the three largest passive firms, hold a combined 25% of FactSet’s shares. Among the top 20 shareholders, I suspect only Baron Capital, a 10.2% shareholder, led by one of the all-time investment greats, Ron Baron, is active.

Altogether, the top 20 shareholders were net buyers (2.6 million shares or 7% of shares outstanding) of FactSet’s shares during Q4 2025 from the latest 13F filings.

In late December 2025, Ron Baron spoke quite highly of FactSet and its new CEO during a CNBC interview, and in the last five months, Baron Capital have increased its position by 108%, the first substantial increase since Q1 2008, having held its shares for more than 20 years. This is also the first time they’ve ever held more than 10% of FactSet’s shares.

As you can imagine, Baron Capital would likely have a key say on if FactSet gets sold (they also own shares in the other two options, Morningstar and Gartner), and given the history and profits gained over the decades investing in FactSet, if I would guesstimate, I suspect they would probably push for a 40% premium to today’s price, $297 per share or a $11 billion market cap exit.

Final thoughts

While I’d prefer for FactSet to stay public, after weighing up the broader competitive landscape, the stated PE interest and some speculative thoughts, I suspect FactSet might not be a publicly listed company by the end of the year.

For now, I’m staying put, though. I maintain the full FactSet investment position and continue to monitor developments in the broader AI landscape, its peers and of course, FactSet’s earnings results (Q2 2026 FY), next scheduled for 31st March 2026.

As a long-term investor, I do hope this is all speculation, but I will double down on researching new ideas in case of a take-private offer.

Subscribe now

After spending most of my Q1 2026 studying the global payment networks, I concluded on the price range I’d pay today for each of the three networks:

• Visa (25x my 2026 estimated net earnings)

• Mastercard (26x my 2026 estimated net earnings)

• American Express (18x my 2026 estimated net earnings)

Combined with EPS growth and dividends, I concluded that buying at the purchase multiples above could meet a minimum 5-year IRR of 15% (or a double in 5 years).

At the start of the research, only Visa met this target, but during the research, Mastercard joined the list through a combination of its share price decline and my improved understanding of its growth potential. Combined, both Mastercard and Visa currently make up 13% of the Jenga Global Fund.

You can read the Visa deep dive here.

At the start of the year, American Express, then a 25x P/E stock, well-above the 18x 2026E purchase multiple I assigned, was far off the buy range and given the then sentiment for its shares, it seemed likely to remain elevated.

Unexpectedly, the sentiment has quickly shifted, and its shares have declined by -21% in two months and now sit within my minimum 15% IRR target.

So what am I doing with this information?

Portfolio management

For now, nothing.

• While I’m optimistic about the resilience and prosperity of the payment networks, making them more than 15% of the Fund also wouldn’t be prudent from a portfolio risk management lens.

• American Express requires more margin of safety due to its higher exposure to the consumer credit cycle, unlike Visa and Mastercard.

• I’m also not a fan of too much portfolio activity, and at current prices, the risk/reward between Visa/Mastercard and Amex isn’t wide enough to justify reducing shares in the former to include Amex in the portfolio.

That said, there’s a price and an IRR difference I concluded justifies making changes here, a 10% IRR difference.

An example is if Visa’s shares rise into a 10% IRR potential ($404 per share, from $314 today), while Amex declines even further from today into a 20% IRR potential ($259 per share, from $303 today), thus leading to a 10% IRR difference between Visa and Amex.

Portfolios always face opportunity costs, and over time, I’ve come to learn the importance of factoring this cost when making investment and portfolio decisions.

So for now, I’ll be maintaining the current portfolio weighting while spending more time learning even more about American Express and stress testing current business and earnings growth assumptions.