By mid-May 2026, Adobe was closer to $100 billion.

So roughly $225-$230 billion of market value has disappeared. That is about a 70% decline from the peak.

Back then, Adobe was treated like one of the cleanest software compounders in the world: high recurring revenue, high gross margin, strong free cash flow, and deep creative workflows. Today, the question is whether generative AI lowers the value of basic creation, shifts workflows toward other AI platforms, raises costs, and makes it harder for Adobe to turn usage into paid recurring revenue.

Adobe is still a very profitable software business with a strong five-year record, but AI is attacking the exact area where Adobe has historically made money.

What Adobe Really Is

The company makes money mainly through subscriptions. In fiscal 2025, Adobe generated $23.8 billion of revenue. Of that, $22.9 billion was subscription revenue. It is collecting recurring revenue from tools and platforms that people and companies use repeatedly.

Digital Media is the main business. In FY2025, it produced $17.65 billion of revenue, grew 11%, and generated $16.81 billion of gross profit. That is where Creative Cloud, Document Cloud, Acrobat, Photoshop, Illustrator, Premiere, Express, Firefly and the related tools sit.

Digital Experience is smaller, but still meaningful. It produced $5.86 billion of revenue, grew 9%, and generated $4.24 billion of gross profit. This is where Experience Cloud, Adobe Experience Platform, GenStudio, analytics, customer data, marketing workflows, commerce, Workfront and enterprise experience software sit.

Publishing and Advertising is no longer central to the thesis. It produced only $0.26 billion of revenue in FY2025, declined 7%, and generated $0.17 billion of gross profit.

The core software economics are still very strong. Digital Media produced gross profit equal to 95% of revenue. That does not mean 95% drops to the bottom line, because Adobe still spends heavily on R&D, sales, marketing and administration. But it does show why the business has been such a strong cash generator.

The AI debate is about protecting this high-quality base while building new revenue streams. If AI helps Adobe sell more Creative Cloud, Acrobat, Firefly, Express, GenStudio and Experience Platform products, the business can keep compounding. If AI forces Adobe to give away more value for free or accept lower pricing, the historical economics become less useful.

In fiscal 2026, Adobe changed the reporting structure and now has one operating and reportable segment. From an investor perspective, that gives less visibility, although Adobe still gives customer group subscription revenue.

For the full fiscal year 2025, Creative & Marketing Professionals produced $16.30 billion of subscription revenue and grew 11%. Business Professionals & Consumers produced $6.50 billion and grew 15%. Together, Digital Media and Digital Experience subscription revenue was $22.80 billion, up 12%.

The same pattern continued in Q1 FY2026. Creative & Marketing Professionals produced $4.39 billion of subscription revenue and grew 12%. Business Professionals & Consumers produced $1.78 billion and grew 16%. Total Customer Group subscription revenue was $6.17 billion, up 13%.

Creative & Marketing Professionals is still the larger revenue base. Business Professionals & Consumers is smaller, but it was the faster-growing customer group in FY2025 and again in Q1 FY2026.

That helps explain the strategy. Adobe is not just a creative professional company anymore. It wants to serve professional creators, marketers, office workers, consumers, students, small businesses and large enterprises.

If AI makes creation easier, more people can create. A student can make a presentation. A small business owner can make an advert. A marketer can generate campaign variations. A large company can automate thousands of content assets. A normal office worker can turn a PDF into a summary, deck, infographic or audio file.

The Products

1. Creative Cloud

Creative Cloud is still the heart of the company. It includes products like Photoshop, Illustrator, Premiere Pro, After Effects, Lightroom and InDesign. These are the tools many professional creators, designers, video editors, photographers and creative teams already know.

Adobe said FY2025 Digital Media subscription growth was driven by strength in Creative Cloud Pro, other flagship apps and Acrobat. In Q1 FY2026, management again said Creative Cloud growth was driven by the CC Pro offering.

This is the core cash engine. It is not only because the apps are good. It is because the products are learned, shared, used inside teams, connected to files, tied to creative habits, and used in professional workflows. That is harder to replace than a single image-generation tool.

The risk is that AI makes some parts of creation easier and cheaper. If that only helps Adobe users do more inside Creative Cloud, it is positive. If it makes customers question why they need the full paid workflow, it becomes a problem.

2. Acrobat, Acrobat Studio And Express

Acrobat is the document side of Adobe. It lets users read, edit, sign, create and work with PDFs. That may sound less exciting than AI image generation, but documents are a huge daily workflow. Businesses do not stop using PDFs because a new AI tool appears.

Adobe is now trying to turn Acrobat from a document tool into a broader productivity and creation platform. Acrobat Studio brings together Acrobat, Express and AI agents. Acrobat AI Assistant can summarise documents, answer questions, cite sources and help create formatted content.

That group produced $6.50 billion of subscription revenue in FY2025, up 15%, and $1.78 billion in Q1 FY2026, up 16%.

The latest product signals are good. Acrobat + Express MAU grew about 20% year over year in Q1. Acrobat AI Assistant ARR grew about 3x. Management also said Acrobat Studio saw strong upgrades as part of enterprise licence renewals.

3. Firefly

Firefly is Adobe’s generative AI product family. It is not one small feature. It includes Adobe’s own Firefly models, the Firefly app, Firefly credit packs, Firefly Enterprise, Firefly Services and Firefly Foundry.

Firefly helps users generate and edit creative content with AI. It is used across images, design, video and other creative workflows. Adobe’s own filings describe Firefly models as commercially safe and trained on licensed content and public domain assets. That is important for enterprise customers because companies care about rights, brand safety and legal risk.

Firefly is Adobe’s clearest AI revenue signal. Adobe is not only adding AI features to defend the old business. It is trying to charge for AI through the Firefly app, credit packs and enterprise products.

Firefly ending ARR exceeded $250 million in Q1 FY2026. Total Adobe ARR was $26.06 billion. Firefly is real, but still small. It proves Adobe can charge for AI, but it does not yet prove that AI can move the whole company.

4. Adobe Stock

Stock is the warning sign.

Adobe Stock gives customers access to licensed creative assets. Historically, that was a useful model. If you needed an image, video or other asset, you could license it rather than create it from scratch.

Generative AI changes that behaviour. If a user can generate a good enough asset, the old standalone Stock model weakens.

Management said the traditional standalone Stock business declined more than expected in Q1 FY2026 and that the shift is happening faster than planned. Adobe’s strategy is to give customers a choice between stock and generative AI for creative and marketing workflows.

I like that they are being direct about the issue. But I would not ignore it. Stock may be small compared with the whole company, but it is one of the cleanest examples of AI replacing an older content model.

5. GenStudio And Adobe Experience Platform

GenStudio and Adobe Experience Platform sit more on the enterprise marketing side.

Adobe describes GenStudio as a content supply chain solution. In plain English, it helps large companies plan, create, produce, activate, deliver, manage and measure marketing content. It connects tools like Adobe Experience Manager Assets, Workfront, Firefly Services and GenStudio for Performance Marketing.

Adobe Experience Platform is the customer data and orchestration layer. It helps companies collect and activate customer data across channels, then use that data for personalization, journey management, analytics and marketing execution.

This is not as easy to understand as Photoshop or Acrobat, but it may be very important. Large companies need more content, more personalization, more measurement and more control. If AI increases content volume, enterprises need systems to manage that content without losing brand control.

Management said GenStudio and AEP & Apps ending ARR each grew over 30% year over year in Q1 FY2026. It also said the number of customers with ARR over $10 million grew more than 20%.

Enterprise demand is still there. The issue is whether it becomes a larger profit engine over time. Digital Experience had $5.86 billion of revenue in FY2025 and $4.24 billion of gross profit. That is a good business, but the gross profit rate is lower than Digital Media. It needs scale and execution.

6. Workfront, Frame.io, Substance 3D And Other Workflow Products

Adobe also owns products that are less visible to casual investors but matter inside workflows.

Workfront is work management for teams. Frame.io supports video collaboration and review. Substance 3D is used for 3D content creation. Adobe Commerce, Marketo Engage, Journey Optimizer, Analytics, Customer Journey Analytics and Real-Time CDP all sit inside the enterprise experience stack.

Adobe does not disclose each product’s revenue or profit. These products matter because they make the platform harder to replace. The more Adobe connects creation, collaboration, data and activation, the stronger the workflow argument becomes.

The main enterprise question is whether GenStudio, AEP and Experience Cloud can become a larger part of the cash flow story.

The Five-Year Record

Adobe’s recent numbers still show a business that has grown. Revenue went from $15.8 billion in fiscal 2021 to $23.8 billion in fiscal 2025.

Revenue was $15.8 billion in FY2021, $17.6 billion in FY2022, $19.4 billion in FY2023, $21.5 billion in FY2024, and $23.8 billion in FY2025.

Operating cash flow also moved higher over time. It was about $7.2 billion in FY2021, $7.8 billion in FY2022, $7.3 billion in FY2023, $8.1 billion in FY2024, and $10.0 billion in FY2025.

That is roughly 11% annual revenue growth over four years. The cash flow record is also strong, although not perfectly smooth. Fiscal 2023 and fiscal 2024 were not as clean as fiscal 2025. But by fiscal 2025, Adobe produced more than $10 billion of operating cash flow.

Capex is very low for a company of this size. In fiscal 2025, Adobe spent only $179 million on property and equipment. That leaves roughly $9.85 billion of free cash flow if I use operating cash flow minus capex.

For an owner, the point is that Adobe does not need a lot of physical capital to keep the business running. The company spends heavily on people, R&D, sales, marketing, cloud infrastructure and products, but the cash economics are still very good.

In FY2025, GAAP operating income was $8.7 billion. Non-GAAP operating income was about $11.0 billion. In Q1 FY2026, GAAP operating income was $2.4 billion and non-GAAP operating income was $3.0 billion. Q1 non-GAAP operating income was 47.4% of revenue.

The Business Engine

First, the tools are deeply used. Creative professionals learn Photoshop, Illustrator, Premiere, After Effects, InDesign and Lightroom. Companies build teams, templates, libraries, brand processes and production workflows around these products. In documents, Acrobat and PDF have decades of habit behind them. In marketing, Adobe is used inside large companies that need analytics, content management, customer data, campaign management and workflow control.

Second, the business is recurring. Customers do not buy a new Photoshop box every few years. They subscribe. Enterprises renew. Users upgrade. Teams add seats. Customers move between plans and products.

Third, Adobe is trying to attach AI to workflows customers already use. A prompt-based AI image tool may be useful, but it is not the same thing as a full production workflow. A brand team does not only need an image. It needs rights, consistency, editing, approvals, brand safety, collaboration, formatting, storage, delivery and measurement.

Q1 FY2026

Q1 FY2026 had strong operating numbers and early AI signals. Total revenue was $6.40 billion, up 12%. Subscription revenue was $6.20 billion, up 13%. Total Adobe ARR was $26.06 billion, up 10.9%. Operating cash flow was $2.96 billion, and Adobe repurchased about 8.1 million shares in the quarter.

The usage numbers were also strong. Adobe surpassed 850 million monthly active users across Acrobat, Creative Cloud, Express and Firefly, up 17%. Creative freemium MAU crossed 80 million, up more than 50%. Acrobat + Express MAU grew about 20%.

The AI monetization signs are still early, but they are not zero. AI-first application ARR more than tripled year over year. Firefly ending ARR, across Firefly App, Firefly credit packs and Firefly Enterprise, exceeded $250 million. AEP & Apps and GenStudio ending ARR each grew more than 30%, and total customers with ARR over $10 million grew more than 20%.

The latest quarter shows why the AI debate is difficult. User growth is moving first. MAU, freemium usage, credit consumption, Express, Firefly and Acrobat AI Assistant are all moving in the right direction.

The revenue proof is earlier. Firefly ending ARR is above $250 million, while total Adobe ARR is about $26 billion. Adobe needs more than usage. It needs paid conversion and durable economics.

The Stock Business Is A Warning

Management said the traditional standalone Stock business declined more than expected in Q1, and that the shift is happening faster than planned.

I would not make too much of this by itself. Stock is not the whole Adobe business. But it is still useful evidence because it shows where AI can substitute for an older content model.

If a customer can generate an image instead of buying a stock asset, part of the old model weakens. Adobe knows this and is trying to give customers a choice between stock and generative AI. That may be the right strategy. But it still shows the risk.

The bigger question is whether Stock is the small part that gets disrupted, or the early example of a wider change in creative workflows.

If the disruption stays around standalone Stock and Adobe replaces it with Firefly, credit packs, Firefly Enterprise and GenStudio, the bear case weakens.

If the same pressure spreads into Creative Cloud pricing, upgrade behaviour or enterprise renewals, the bear case gets much stronger.

The Bull Case

The bull case is not that AI does not matter. The bull case is that AI makes Adobe more useful.

A person who never learned professional software may start with Express, Firefly or Acrobat AI Assistant. Some of those users may never pay. But some may. A student may become a creator. A small business may become a paid team. An enterprise may use Firefly Services to automate content production. A brand may use GenStudio to generate, review, govern and activate assets across channels.

Adobe’s advantage is that it can connect creation with editing, documents, brand control, customer data, marketing workflows and enterprise governance.

A generic AI model can create something impressive. But large companies need more than impressive. They need content that fits brand rules, legal requirements, approval processes and campaign systems. They need to know where the asset came from. They need controls. They need reliability. They need people across teams to work with the same files and the same standards.

Adobe’s old strengths may become useful again when enterprises need governed content, approvals and measurement, not only a generated asset.

The company has products across the whole chain:

• Photoshop, Illustrator, Premiere, After Effects, Lightroom and InDesign for professional creative work.

• Acrobat, PDF, Acrobat Studio and AI Assistant for document work.

• Express and Firefly for lighter creation and new users.

• Firefly Services and Firefly Foundry for enterprise AI content production and custom models.

• Frame.io and Workfront for collaboration and production workflow.

• Adobe Experience Manager, Adobe Experience Platform, Analytics, Journey Optimizer, Commerce, GenStudio and related products for customer experience and marketing workflows.

I would not describe Adobe as only a tool company. The stronger version is that Adobe owns important parts of the content and marketing workflow.

The Bear Case

The bear case is also real. It is not that Adobe disappears. It is that the business becomes a slower-growing, more expensive-to-run software company with less pricing power. The risk shows up in seven places.

First, AI may reduce the value of basic creation. If a user can create a good enough image, video, presentation or design in a chatbot or a cheap AI app, Adobe has to prove why the paid workflow is still worth it. Professionals may stay. Enterprises may stay. But entry-level and casual users may be harder to monetize.

Second, Adobe may get usage without economics. Management is clear that freemium is part of the strategy. Creative freemium MAU grew more than 50%, which sounds excellent. But management also said this approach has a near-term impact on ARR. That is the right trade-off if free users convert later. It is the wrong trade-off if Adobe trains users to expect more creation for free.

Third, AI may shift distribution away from Adobe-owned surfaces. Adobe is integrating with ChatGPT and other AI platforms. That can bring reach. But it can also make Adobe more dependent on third-party platforms. If users start creation inside OpenAI, Google, Microsoft, Anthropic or another assistant, Adobe needs to make sure it still owns enough of the workflow and economics.

Fourth, AI can raise costs. Adobe’s own 10-Q says developing, testing and deploying AI systems can increase compute costs and that adoption and monetization pathways are uncertain. That is plain enough. AI can make the product better, but if the company must spend more on infrastructure, models, datasets, compliance and indemnity, the cash flow argument becomes more complicated.

Fifth, AI brings legal, IP and regulatory risk. Adobe’s filings discuss AI regulation, third-party model risk, training data, IP claims, privacy claims and reputation risk. Adobe sells trust to enterprises. If customers are buying Adobe partly because it is commercially safer, Adobe has to keep that trust.

Sixth, reporting is now less detailed. Adobe combined its old segments into one operating and reportable segment in Q1 FY2026. Management says this reflects unified selling and integrated product innovation. That may be true. But from an investor view, less segment detail makes it harder to see exactly what is working and what is weakening.

Seventh, there is a CEO transition. Shantanu Narayen has led Adobe for a long time.

What The Numbers Say Today

Q1 FY2026 revenue grew 12%. Customer group subscription revenue grew 13%. Operating cash flow was a Q1 record at $2.96 billion. The company had $6.89 billion of cash and short-term investments exiting the quarter. Non-GAAP operating income was 47.4% of revenue.

For FY2026, management reaffirmed revenue of $25.90 billion to $26.10 billion. It expects Business Professionals & Consumers subscription revenue of $7.35 billion to $7.40 billion, Creative & Marketing Professionals subscription revenue of $17.75 billion to $17.90 billion, Total Adobe ending ARR growth of 10.2%, and non-GAAP operating margin around 45%.

If Adobe can grow revenue around 9-10%, keep non-GAAP profitability around the mid-40s as a percentage of sales, and keep producing close to $10 billion or more of free cash flow, the market does not need a perfect AI story.

By mid-May 2026, Adobe was roughly a $95-$100 billion company. Against FY2025 free cash flow of about $9.85 billion, that is around 10x owner earnings. Against my rough FY2026 cash generation base of about $10.5 billion, it is closer to 9-10x. The stock is no longer priced like a perfect software story.

Valuation

Before I put numbers into a DCF, I want to think about the stock like I am buying the whole business.

That is the Buffett way of thinking I find useful. It is a piece of a business. The market cap is what the market is asking me to pay for the whole company today. The value of the business is the present value of the cash it can generate for owners over its life.

So the question is not only whether Adobe trades at 9x, 10x or 12x free cash flow.

The question is whether the whole business is worth more than the market cap today, after discounting the future cash Adobe can realistically produce.

In FY2025, Adobe produced $10.03 billion of operating cash flow and spent $179 million on property and equipment. That gives me about $9.85 billion of free cash flow.

For FY2026, management is guiding to $25.90-$26.10 billion of revenue, 10.2% Total Adobe ending ARR growth, and non-GAAP operating margin around 45%. Using the midpoint of the revenue guide and a low-40s free cash flow margin, my working free cash flow base is about $10.5 billion.

If AI weakens pricing power and Adobe becomes a no-growth cash flow business, I would use a low terminal multiple, maybe 8-10x free cash flow. If Adobe stays durable but grows more slowly, 12-15x free cash flow is more reasonable. If AI expands the workflow and Adobe keeps strong growth and high margins, then 18x can be argued. I would be careful using anything above that unless the evidence becomes much clearer.

In the stress case, I assume free cash flow declines 2% a year, use an 11% discount rate, and apply an 8x terminal multiple to year-10 FCF. That gives about $81 billion of equity value, or roughly $201 per share.

In the no-growth case, I assume cash generation stays flat and apply the same 8x terminal multiple. That gives about $98 billion of equity value, or roughly $241 per share.

In the bear-but-still-durable case, I assume free cash flow grows 3% a year and the terminal multiple is 10x. That gives about $130 billion of equity value, or roughly $320 per share.

In the base case, I assume owner cash grows 6% a year and Adobe still deserves a 15x terminal multiple. That gives about $196 billion of equity value, or roughly $484 per share.

In the bull case, I assume free cash flow grows 9% a year and the terminal multiple is 18x. That gives about $273 billion of equity value, or roughly $676 per share.

I do not treat this as a target price. I use it to see what the current price already assumes.

At around a $95-$100 billion market cap, the market is close to my no-growth case. That means investors are not paying much for future growth today.

If the bear case is right and AI slowly reduces Adobe’s pricing power, raises costs, weakens Creative Cloud monetization, and turns more usage into lower-quality revenue, then the stress case matters. A low multiple will not protect me if the cash flow base starts going backwards.

But if Adobe can keep producing more than $10 billion of free cash flow, grow that cash flow even in the low single digits, and still deserve a 10-12x terminal free cash flow multiple, the current valuation looks too cheap.

Adobe has to show that AI is not only defending the old business. It has to become a paid workflow layer across Creative Cloud, Acrobat, Express, Firefly, GenStudio and Experience Cloud. That is what would support the base or bull cases.

In FY2025, the company repurchased about 30.8 million shares and spent $11.3 billion on common stock repurchases. In Q1 FY2026, it repurchased another 8.1 million shares for about $2.5 billion. The March 2024 authorization was for up to $25 billion, and $3.89 billion remained at the end of Q1.

My Current View

The current business is strong. The five-year record is good. Adobe still produces a lot of cash. The products are deeply used. The balance sheet is fine. The buyback can help per-share value.

But the bear case is not imaginary. AI attacks the centre of the company. It changes who can create, where they create, how much they pay, what tools they need, and what platforms own the customer relationship.

Adobe has a real chance to use AI to expand the number of creators and business users. More people can create because the tools are easier. More companies need content because digital channels keep multiplying. More marketing teams need automation because manual content production cannot keep up. If Adobe keeps the workflow, trust and cash economics, AI can make the company more valuable.

The valuation is what makes the idea interesting. At around a $95-$100 billion market cap, Adobe is close to my no-growth case. My stress case was around $81 billion of equity value. Based on my valuation work, I think downside risk is more limited.

I want ARR to stay around double-digit growth while free cash flow remains strong. I want Firefly, Acrobat AI Assistant, GenStudio, AEP AI products and Firefly Enterprise to become material revenue sources, not only usage stories. I want Creative Cloud and Acrobat upgrades to show that customers will pay more because AI makes the products more valuable. I also want margins and cash conversion to hold while Adobe invests in AI.

Stock is the product-level warning sign. If standalone Stock declines but Adobe replaces it with Firefly and enterprise content automation, that is acceptable. If Stock is the first sign of broader Creative Cloud pricing pressure, the view has to change.

Enterprise retention, Creative Cloud pricing and seat behaviour are the faster signals. If customers start using AI to reduce Adobe seats rather than expand usage, the thesis changes quickly. Cost matters too. AI compute, model costs, compliance, indemnity, data and infrastructure spending cannot keep rising if cash generation starts to weaken.

Adobe is going through a major product transition at the same time as a leadership transition. That does not have to be a problem, but it is not nothing.

The question is whether AI increases Adobe’s useful role in the workflow, or slowly gives away the value of creation to cheaper tools and third-party platforms.

Not investment advice. Just my own thinking.

E-commerce, social media, digital payments, delivery, games, and online services were not just nice things to have. In many places, they helped solve problems where infrastructure was still catching up.

That created a big tailwind for companies like Alibaba ($BABA), Tencent ($TCEHY) , Baidu ($BIDU), JD.com ($JD), Meituan (3690.HK), and NetEase ($NTES).

For a long time, these companies were growing revenue, entering new markets, and becoming more important in everyday life, so the market was willing to pay for that growth.

After 2020, the story changed. Regulation became more important, the internet crackdown hurt investor confidence, consumer behaviour changed, and the macro economy in China became harder. The property market also became a major pressure point. Property activity fell sharply: real estate development investment fell from RMB14.1 trillion in 2020 to RMB8.3 trillion in 2025, while commercial housing sales value fell from RMB17.4 trillion to RMB8.4 trillion.

That changed the background for Chinese consumer and internet companies.

This is not only a company-specific story. It is part of a wider reset in how investors think about Chinese internet companies.

Tencent’s market cap at HKD4.298 trillion, which is about US$550 billion. Alibaba is more famous in the West because of Jack Ma and because e-commerce is easier to explain. Tencent is less familiar, but I think it is the more interesting business to study because it touches so many parts of digital life in China.

Business Quality

Tencent is not only a gaming company. It owns several different businesses: domestic games, international games, social networks, marketing services, FinTech, business services, cloud, and Weixin/WeChat.

In 2025, Tencent reported RMB751.8 billion of revenue, up 14%.

• VAS: RMB369.3 billion, up 16%, about 49% of revenue

• Marketing Services: RMB145.0 billion, up 19%, about 19% of revenue

• FinTech and Business Services: RMB229.4 billion, up 8%, about 31% of revenue

• Others: RMB8.1 billion, about 1% of revenue

VAS includes domestic games, international games, and social networks. In 2025, domestic games revenue was RMB164.2 billion, international games revenue was RMB77.4 billion, and social networks revenue was RMB127.7 billion.

The part I care about most is not only the segment split. It is the platform position. Weixin (WeChat) had 1.418 billion combined monthly active users at the end of 2025.

For many users, WeChat is not just an app. It is where they message, pay, use services, interact with merchants, use mini programs, watch content, and move through large parts of daily digital life.

That makes Tencent different from a company that only competes by cutting price or spending more on customer acquisition. Tencent still has competition, and I do not want to pretend it does not. But many businesses also need to work with the ecosystem. Merchants, advertisers, developers, payment users, game studios, cloud customers, and service providers all touch parts of Tencent.

This is why I see Tencent closer to a digital infrastructure business with several profit pools attached to it.

Growth Drivers

Tencent has several very profitable segments, and many of them still have growth tailwinds. Games are still the main profit engine. Advertising is getting better. Payments and fintech continue to grow. Cloud looks more disciplined than before. Mini Programs make the WeChat ecosystem stronger.

Marketing Services can grow if user engagement stays strong in Video Accounts, Weixin Search, Mini Programs, Mini Shops, and better ad targeting. FinTech and Business Services can grow through commercial payments, wealth management services, consumer loan services, cloud services, AI-related cloud demand, and e-commerce technology service fees.

Games can grow through both domestic and international games, but gaming always has content-cycle and regulatory risk.

The other growth driver is Tencent’s ecosystem and investment muscle. Tencent has invested in public and private companies for many years, and in some cases those companies also connect back to Tencent’s ecosystem, but I would separate that value from Tencent’s operating growth.

Kaspi.kz is one example. Kaspi ($KSPI) announced in April 2026 that Tencent invested alongside Kaspi’s co-founder and other long-term investors in the purchase of 6.0 million ADSs from Baring Fintech Venture Funds. Kaspi called Tencent one of its largest shareholders.

PDD is another useful example. Tencent is still a major shareholder in PDD, owning 13.8% of PDD’s Class A ordinary shares through entities affiliated with Tencent. The two companies have also had a strategic cooperation relationship since 2018, including access points on Weixin Pay, payment solutions, cloud services, user engagement, Weixin payment services, and technical support. PDD also said that in 2023 it signed new agreements with Tencent for continued access points on the Weixin platform.

Tencent is not only an operating company. It also has an ecosystem and portfolio layer around other digital businesses.

Margin / Free Cash Flow Drivers

The financials are one reason I keep following Tencent quarter by quarter. China has been a difficult market since 2020, but Tencent’s profits and cash flow have held up much better than Alibaba’s and many other large Chinese names.

Between 2021 and 2025, Tencent’s revenue grew from RMB560.1 billion to RMB751.8 billion. But the more interesting part is what happened below the revenue line. Gross profit increased from RMB245.9 billion to RMB422.6 billion.

IFRS operating profit increased from RMB124.7 billion to RMB241.6 billion. Non-IFRS profit attributable to shareholders increased from RMB123.8 billion to RMB259.6 billion.

So Tencent did not just grow. It became more profitable during a very tough period. That matters because the environment was difficult for everyone, including Tencent. Broader economic activity slowed down, and many Chinese internet companies struggled to protect margins and cash flow.

The difference becomes clearer when you compare Tencent with Alibaba.

As I wrote before, Alibaba’s revenue growth became more expensive. Margins were pressured, and free cash flow declined. Tencent was different. Revenue grew, but profits grew faster. Margins improved, and free cash flow stayed strong.

Gross margin improved from around 44% in 2021 to 56% in 2025. IFRS operating margin improved from around 22% to 32%.

That matters because revenue growth alone can be misleading. What I want to see is whether more of each renminbi of revenue is turning into profit. In Tencent’s case, that has clearly happened.

For FY2025, Tencent reported RMB751.8 billion of revenue and RMB422.6 billion of gross profit. More importantly, the business generated RMB182.6 billion of free cash flow, up 18% for the year. Tencent also ended the year with RMB107.1 billion of net cash.

Capital Allocation

Capital allocation is one of the main reasons Tencent is interesting. I do not want to overdo the Berkshire ($BRK.B) comparison, but there is one simple similarity: Tencent has used operating cash flow to own stakes in many other companies over time.

At the end of 2025, Tencent reported RMB672.7 billion of listed investee company shareholdings at fair value and RMB363.1 billion of unlisted investee company shareholdings at carrying value.

Tencent also repurchased 153.4 million shares for about HKD80.0 billion in FY2025 while ending the year with RMB107.1 billion of net cash.

That combination matters to me because the company can reinvest in games, cloud, AI, payments, and WeChat, while also buying back shares and keeping financial strength.

The shareholder structure is also worth noting. As of 31 December 2025, MIH Internet Holdings B.V., controlled by Naspers through Prosus, owned 22.8% of Tencent. Ma Huateng, Tencent’s co-founder, had an 8.82% interest through Advance Data Services.

Market Concern

The market has real reasons to be cautious. Tencent is Chinese, complicated, and exposed to regulation, gaming approvals, geopolitics, currency, consumer weakness, AI spending, and the wider discount investors apply to Chinese equities.

There is also one honest limitation. I do not live in China. I can read the annual reports and filings, but that is not the same as using the ecosystem every day. Tencent is tied to how people in China communicate, pay, shop, play games, consume media, and use services. It is also tied to local regulation. Because of that, I need a bigger margin of safety than I would need for a business I understand directly as a customer.

Bear Case

The bear case is not that Tencent is a bad business. The bear case is that Tencent can be a very good business and still not be a good investment for a foreign investor if the discount is not large enough.

Regulation can change. Gaming approvals can affect the profit engine. Consumer behaviour can weaken. AI and cloud spending can pressure free cash flow if returns are not good enough. The investment portfolio can be hard to value, especially the unlisted part. Minority shareholders may not receive the full value of the portfolio in the simple way a spreadsheet suggests.

There is also the risk that I overestimate the ecosystem because I am looking from outside China. WeChat may be critical to daily life, but I still need to understand what can weaken that position, what can become more competitive, and what the government may allow Tencent to earn over time.

This is why the short thesis matters more than the long thesis at the start. It is easy to say Tencent is high quality. The harder question is why I might be wrong.

Valuation Question

As of today, Tencent trades at HKD471.40, with a market value of around HKD4.3 trillion and a P/E of about 17x.

Tencent trades at around 20–21x FY2025 free cash flow based on the headline market cap. That is not deep value by itself.

The valuation becomes more interesting when I adjust for the balance sheet and investment portfolio. Tencent had RMB107.1 billion of net cash and RMB672.7 billion of listed investee stakes at fair value. If I give credit to those two items, the operating business looks closer to a mid-teens multiple of FY2025 free cash flow.

If I also give full credit to the unlisted investments, the multiple looks even lower. But I would be careful with that. Unlisted investments are not cash. I do not want the thesis to depend on treating them as cash.

So the valuation question for me is simple:

After cash, listed investments, the broader portfolio, and China risk, is the price attractive enough for the quality of the business?

What Must Remain True

Tencent is already part of the infrastructure of daily life in China.

Weixin and WeChat are not just apps. People use them to communicate, pay, shop, access services, follow content, play games, interact with merchants, and move through the digital economy. That is what makes Tencent different. It is not just another internet company trying to compete on price.

This is why I am less worried about normal competition. Competition always exists, but Tencent’s position is very hard to replicate. The bigger risks for me are regulation, China risk, and how much profit Tencent is allowed to earn from such an important position over time.

The business already has very profitable segments. Games, advertising, payments, fintech, cloud, and business services all sit inside a very strong ecosystem. The important thing is not whether Tencent suddenly becomes profitable. It already is. The question is whether those economics remain strong and whether free cash flow continues to hold up.

Management is also one of the reasons Tencent is interesting. This is not a weak management team with no history. Tencent has a strong track record, highly capable leadership, meaningful insider ownership, and a long-term way of thinking. Ma Huateng still has a serious economic interest in the business, and that matters to me. It does not guarantee perfect decisions, but it changes how I read capital allocation.

I want management to keep doing what they have mostly done well: allocate capital rationally, protect the strength of the core business, invest where returns make sense, and return capital when the stock price is attractive.

I do not want to own Tencent only because the stock is below its old high. I would only want to own it if the business quality, cash generation, balance sheet, investment portfolio, management quality, and price together compensate for the risks.

What Would Prove Me Wrong

I would be wrong if Weixin started to lose their role as daily digital infrastructure in China.

The real risk would be deeper than normal competition: people using the ecosystem less, merchants becoming less dependent on it, Tencent losing pricing power, or the company losing the ability to monetise its position. As of today, I think that outcome is very unlikely.

I would also be wrong if regulation permanently damaged the economics of the business. The government can change rules, gaming approvals can affect the profit engine, and policy can limit how much profit Tencent is allowed to earn over time.

Competition is still worth watching, but I am less worried about competition alone because Tencent’s position is already very strong. What would worry me more is competition plus regulation plus changing user behaviour weakening the platform over time.

I would also be wrong if margins reversed without a clear future return, if free cash flow weakened structurally, if AI and cloud spending consumed too much cash without good returns, or if the investment portfolio could not be monetised in a way that benefits shareholders.

Tencent can be a great business and still become a poor investment if the price does not leave enough margin of safety.

Not investment advice. Just my own thinking.

Free cash flow was healthy. Management also raised full-year guidance.

For most companies, that would be enough.But software is not being valued normally right now.

The quarter was strong across growth, profitability, cash generation, and customer metrics.

Revenue was $881.0 million, up 23% as reported and 18% in constant currency. Subscription revenue was $862.3 million, up 23%.

Profitability also improved. GAAP operating income was $27.9 million, compared with a loss of $27.5 million last year. Non-GAAP operating income was $156.8 million, up from $100.3 million last year. GAAP operating margin improved to 3.2%, compared with negative 3.8% last year, while non-GAAP operating margin improved to 17.8%, compared with 14.0% last year. Non-GAAP net income was $143.0 million.

Cash generation was healthy. Operating cash flow was $198.8 million, and non-GAAP free cash flow was $153.7 million. HubSpot ended the quarter with $1.8 billion of cash, cash equivalents, and investments.

Customer metrics were also solid. HubSpot had 299,458 customers, up 16%. Average subscription revenue per customer was $11,722, up 6%. Calculated billings were $912.3 million, up 19% as reported and 17% in constant currency.

The company also guided for Q2 revenue of $897 million to $898 million, up 18% as reported and 16% in constant currency.

For the full year 2026, HubSpot guided for revenue of $3.700 billion to $3.708 billion, up 18% as reported and 17% in constant currency.

The company also guided for non-GAAP operating income of $762 million to $766 million, with a non-GAAP operating margin of 21%.

Non-GAAP EPS is expected to be between $13.04 and $13.12.

This is why the stock reaction surprised me.

The stock fell sharply after hours, trading around or below $190. That puts the market cap around $10 billion.

It reported growth, margin expansion, free cash flow, and higher guidance.

So why did the stock fall?

Because investors are no longer giving SaaS companies the benefit of the doubt.

The AI Problem

The concern is not really Q1.

The concern is the future shape of software.

When investors say “AI risk” for software companies, they usually mean tools like ChatGPT, Gemini, Claude, and the new coding tools that make it easier to create software, automate workflows, and interact with data.

For years, software companies benefited from a simple setup:

Companies needed tools. Employees used the tools. The vendor charged per seat, per module, or per platform.

AI makes that model less obvious.

If an AI agent can do the work, maybe the human does not need to log into the software as often. If the AI can update records, write follow-ups, summarize meetings, create tickets, qualify leads, and answer customer questions, then the old software interface becomes less important.

That is the fear.

If AI really changes how work gets done, then a lot of software pricing models will need to change.

The question is whether HubSpot is hurt by that change or whether HubSpot becomes one of the platforms where that change happens.

Agentic Is Easy To Say

This is where I think investors need to be careful.

It is easy to say “agentic but It is much harder to create useful agentic workflows for real businesses. HubSpot has almost 300,000 customers. Those customers do not all use HubSpot in the same way. Some use it mainly for marketing. Some use it for sales. Some use it for service. Some use it as a lightweight CRM. Some use multiple hubs together.

Each business has its own data, workflows, customer history, sales process, support rules, permissions, integrations, and internal habits.

That is the hard part.

AI by itself is not enough.

The AI needs context.

It needs to know the customer, the company, the workflow, the prior conversations, the open deals, the support history, the marketing campaign, the permissions, and the next best action.

That is why I do not think most companies will simply build their own CRM or marketing system from scratch.

Some companies can do it.

I listened to Palantir’s earnings call a few days ago, and they actually talked about legacy SaaS. They also seem to have built their own AI systems in a way that reduces the need for a traditional CRM.

A company like Palantir may be able to build very specific internal AI workflows because it has the technical talent, the operating culture, and the need. In that kind of company, the old CRM idea may disappear. The AI can capture the data, update the system, route the work, and bring the human in only when judgment is needed.

But most companies are not Palantir.

They do not have the time, resources, engineering depth, or internal clarity to build their own CRM, marketing automation, sales automation, and service automation layer.

They still need a platform.

That is HubSpot’s opportunity.

But it is also HubSpot’s risk.

If HubSpot can turn its customer data and workflows into useful agents, it becomes more valuable.

If it cannot, then customers may ask why they are paying for another software layer when AI tools can do more work directly.

The Monetization Question

For me, the biggest question is AI monetization.

HubSpot has been trying to explain this through products like Customer Agent, Prospecting Agent, Data Agent, AEO, Smart Deal Progression, and HubSpot Credits.

I understand the logic.

Instead of charging only for seats, HubSpot can charge for outcomes or usage. For example, Customer Agent is priced around successful resolutions, and Prospecting Agent is priced around recommended leads.

Customers do not want to pay for AI because it sounds interesting. They want to pay because it produces a result.

AI is here, but creating real business use cases is not easy. A company does not commit more budget just because a vendor says “agentic.” It needs to see better conversion, lower support costs, faster ticket resolution, cleaner data, better lead quality, or more efficient sales follow-up.

That takes time.

HubSpot needs customers to use the AI products so the products improve and monetization grows. But customers need to see results before they commit too much spend.

That is why the credit system matters.

If credits become a clean way to charge for real outcomes, then HubSpot has another growth lever.

If credits feel confusing, expensive, or hard to connect to business value, then monetization will be slower.

That is what I am watching.

Capex And Product Investment

The second thing to watch is investment intensity.

HubSpot generated $203.5 million of non-GAAP operating cash flow in Q1 and $153.7 million of non-GAAP free cash flow.

The difference is roughly $50 million.

That came from $15.4 million of purchases of property and equipment and $34.3 million of capitalized software development costs

The company also still has a strong liquidity position. Cash and cash equivalents were $943.9 million, short-term investments were $747.1 million, and long-term investments were $87.5 million at quarter-end. Together, that is about $1.78 billion of cash and investments.

The market is already worried that software companies will need to spend more just to defend their existing business.

That is a very different story from the old SaaS model.

In the old model, investors assumed software companies could scale with very high incremental margins. AI may still improve productivity, but it also requires more product investment, more infrastructure, more data work, and more experimentation.

So the question is simple:

Is HubSpot spending to create a stronger platform, or is it spending just to keep up?

Those are very different things.

Valuation

At around $190 per share, HubSpot’s market cap is roughly $10 billion.

The company has about $1.8 billion in cash, cash equivalents, and investments. That puts the rough enterprise value near $8.2 billion, before smaller adjustments.

Using the midpoint of 2026 guidance, HubSpot expects about $3.704 billion of revenue, about $764 million of non-GAAP operating income, and about $13.08 of non-GAAP EPS.

At $190, the stock trades at roughly 2.7x 2026 revenue, or 2.2x 2026 revenue on an enterprise value basis.

It also trades at about 14.5x 2026 non-GAAP EPS and 10.7x 2026 non-GAAP operating income on an enterprise value basis.

That is not demanding for a business still growing revenue around 18% with a 21% non-GAAP operating margin.

HubSpot is guiding for about $3.704 billion of revenue and $762 million to $766 million of non-GAAP operating income. That is roughly a 21% non-GAAP operating margin.

On that basis, at around $190, the stock trades at roughly 10.7x enterprise value to guided non-GAAP operating income.

Q1 free cash flow is still useful, but as a cash conversion check.

HubSpot reported $153.7 million of non-GAAP free cash flow in Q1.

That number comes from $198.8 million of GAAP operating cash flow, minus $15.4 million of property and equipment purchases, minus $34.3 million of capitalised software development costs, plus $4.7 million of restructuring payments added back.

My View

The market is basically saying: “Good quarter. But prove the AI story.”

Investors are punishing SaaS companies that do not crush earnings and give confidence about the future. Growth deceleration is not a surprise, especially with macro uncertainty. But the old tolerance is gone.

For a few years, SaaS companies could talk about product roadmaps and investors would pay high multiples.

That period is over. Now investors want evidence. They want to see whether AI increases revenue, improves retention, expands margins, or creates new usage-based economics. That is why HubSpot falling after a strong quarter is not completely irrational.

Revenue growth was strong. Margins improved. Guidance moved up. Free cash flow was healthy. Customers continued to grow. Average subscription revenue per customer grew. The balance sheet is strong. The company also bought back $211 million of stock in the quarter, with $789 million remaining under the repurchase program.

But I also understand the concern.

HubSpot has to prove that AI is not just a feature layer. It has to prove that AI can become part of the workflow and part of the monetization model.

Customers need to see results before they commit more spend.

HubSpot has the right ingredients: customer data, workflows, nearly 300,000 customers, multiple hubs, a large ecosystem, and a clear need across marketing, sales, and service.

The stock falling below $190 makes the valuation more interesting. At roughly 10-11x enterprise value to guided non-GAAP operating income, the market is no longer pricing HubSpot like a perfect SaaS compounder.

The question is whether HubSpot can turn AI into measurable business outcomes for customers.

Sources:

- HubSpot Q1 2026 results: https://ir.hubspot.com/news-releases/news-release-details/hubspot-reports-strong-q1-2026-results

- HubSpot Spring 2026 Spotlight: https://www.hubspot.com/spotlight

- HubSpot Credits overview: https://knowledge.hubspot.com/account-management/understand-hubspot-credits-and-billing

We need to separate signal from noise.

Subscribe now

Now Salesforce trades around 13–14x free cash flow. The S&P 500 is closer to 21x. Coca-Cola is around 22x and grows slower. That gap tells you expectations have flipped, the market isn’t paying for the best case anymore. At 14x, Salesforce doesn’t need to prove it can grow at 20% to make sense; it just needs to stay stable, keep customers, and keep converting revenue into cash.

They’re managing costs better

Stock-based comp is now around 8% of revenue, down from over 10%. They also stopped hiring as many people. When costs slow down and revenue stays the same, profits go up.

This is the first time in a while the company seems to care about shareholders.

Customers don’t leave easily

About 90% of the Fortune 500 use Salesforce. It’s not just software. It holds years of customer data, workflows, integrations, and reports. You can’t just turn it off. Replacing it takes years and comes with risk.

So even if growth slows, the customer base stays.

It’s an Ecosystem, Not an App

If Salesforce were just a place to store phone numbers, you could export a file and switch to a competitor in an afternoon. But for the Fortune 500, Salesforce is more like an operating system.

Companies build custom applications on top of Salesforce using proprietary languages like Apex and tools like Flows. These custom apps do not translate to other systems. To leave, you have to rebuild them from scratch.

Subscribe now

A typical enterprise has Salesforce connected to their ERP system, their marketing tools, their customer support lines, and their email. Unplugging Salesforce breaks the data flow across the entire company.

Data Gravity

As a dataset grows, it becomes harder to move.

Years of customer interactions, purchase history, and support tickets are stored there. Migrating this data to a new structure is a large project with a high risk of corruption or loss.

Most companies have messy data. Moving requires cleaning it first—a project that can take months on its own.

The “Rip and Replace” Risk

Replacing a core system like Salesforce is often compared to changing the engine of a plane while it is flying.

If the new system fails or has bugs during the switch, sales stop, support tickets get lost, and revenue takes a direct hit.

You have thousands of employees who know exactly where to click to do their jobs. Moving to a new interface causes a large dip in productivity for 6–12 months as users relearn how to work.

What could go wrong

The stock is cheap for a reason.

1. If teams start working through AI agents, they might not log into Salesforce anymore. The AI just talks to Salesforce in the background. If Salesforce becomes invisible, they lose pricing power.

2. Salesforce charges per user. If AI makes teams more efficient, companies need fewer people. Fewer people means fewer licenses. The question is whether they can replace that revenue fast enough with usage-based pricing.

Why I’m looking at it now

At $175–$190, Salesforce isn’t priced like it has to keep growing fast.

If it becomes like Oracle with customer data, that’s not exciting, but it can generate cash for years. And at ~13x free cash flow, the market already treats it that way.

That’s the point and that’s why it’s worth watching.

Disclaimer: This article is for informational purposes only. It is not financial advice. I am not a financial advisor. I may buy or sell these stocks at any time. You must do your own research before investing

By reading this email, you agree to the full disclaimer found here: https://www.thevaluethesis.com/p/disclamier

I am writing this update to state my position. I am holding the shares.

The thesis shifted. In December, I viewed it as a compounder at a fair price. Today, I view it as a distressed asset.

Here is the analysis of the situation.

Read more

This wasn’t about one issue. Two things hit at the same time.

First: the earnings headline.
UnitedHealth reported very low GAAP earnings and guided to lower revenue next year.

Most trading systems don’t read context. They see “lower revenue” and “no profit” and sell.

Second: the government headline.
The night before earnings, the US government proposed a 0.09% increase in Medicare Advantage rates for 2027.

With inflation running closer to 3–4%, that effectively means getting paid less in real terms.

Put together, the message the market heard was simple:

• Revenue down

• Profits under pressure

• Government pushing back

That combination spooked investors.

What Was Actually Going On in the Quarter

Disclaimer: This article is for informational purposes only. It is not financial advice. I am not a financial advisor. I may buy or sell these stocks at any time. You must do your own research before investing

By reading this email, you agree to the full disclaimer found here: https://www.thevaluethesis.com/p/disclamier

The quarter looked bad on paper, but most of the damage was intentional.

Management took a large one-time charge to deal with existing issues:

• Costs from the cyberattack

• Closing weaker parts of the business

• Cutting underperforming teams and contracts

They chose to deal with these problems now instead of letting them linger.

When you adjust for that, earnings were $2.11 per share, ahead of expectations. The business continued to generate strong cash. The headline numbers just didn’t show it.

Shrinking on Purpose

This is the part that really unsettled the market.

For the first time in years, UnitedHealth is planning to get smaller.

• 2025 revenue: about $448 billion

• 2026 guidance: about $440 billion

This is deliberate.

Management expects to lose:

• Around 1.3–1.4 million Medicare members

• Around 700,000 Medicaid members

These are members that were costing more than they were paying in.

Revenue goes down. Profitability improves.

Instead of chasing size, the company is focusing on margins and control. That usually looks uncomfortable at first.

The Government Question

The biggest unknown is still the government.

Medicare is expensive, and the US is trying to slow cost growth. The message from regulators is essentially:

“We’re not increasing payments. Become more efficient.”

If rates stay tight, insurers reduce extra benefits like dental, vision, and gym memberships. That affects real people.

UnitedHealth covers roughly 8.4 million seniors under Medicare.

Cutting benefits at that scale creates political pressure, especially going into an election year. The final decision hasn’t been made yet.

How AI Fits Into This

AI wasn’t mentioned as a buzzword. It was mentioned as a cost tool.

UnitedHealth plans to spend around $1.5 billion per year on AI, with the goal of saving roughly $1 billion per year starting in 2026.

Where that shows up:

• Call centres: AI helps staff find information faster

• Claims: more decisions handled automatically, with fewer manual steps

If revenue growth slows, lowering costs becomes the main lever. This is how they do that at scale.

Cash Tells a Different Story

While reported earnings were close to zero, cash flow was not.

UnitedHealth generated about $19.7 billion in operating cash flow over the year.

Earnings can be shaped by accounting choices. Cash reflects what the business actually produces.

Management plans to focus on paying down debt in early 2026. After that, they expect to restart share buybacks later in the year.

That matters more than short-term headlines.

Short-Term Issues vs. Long-Term Ones

It helps to separate what’s temporary from what really matters.

Short-term pressures:

• Higher medical usage

• Strong flu seasons

• Temporary cost spikes

These come and go, and are already factored into future pricing.

Long-term pressure:

• Government reimbursement levels

That’s the real debate. It may cap growth and force efficiency, but it doesn’t mean the business stops working.

Management described 2026 as a year to fix things, with expectations of returning to more normal growth in 2027 or 2028.

Looking Ahead: The Possible Paths

Everything from here depends on one thing:

The government’s final Medicare rate decision in April 2026.

From that point, the business likely follows one of three paths.

2026: A Year of Adjustment

This part is fairly clear, based on management guidance.

Revenue comes down slightly as unprofitable members are removed.
Costs improve.
Cash flow remains strong.
Buybacks return in the second half of the year.

Despite lower revenue, earnings still grow because margins improve.

Scenario A

Trigger:
The government raises Medicare Advantage rates to around +1.5% to avoid large benefit cuts for seniors.

Result:
UnitedHealth stops shrinking. Membership stabilises. Share buybacks accelerate in 2027.

Numbers:

• 2027 revenue: approximately $465 billion

• 2027 earnings per share: approximately $20.50

Stock implication:
At a normal valuation, the stock could recover to $430 or higher by 2027.

Scenario B

The government makes a small concession, around +0.5%.

Result:
UnitedHealth continues to shrink slightly for another year but protects margins through cost controls and AI efficiency.

Numbers:

• 2027 revenue: approximately $450 billion

• 2027 earnings per share: approximately $19.75, largely driven by buybacks

Stock implication:
The stock likely moves gradually higher into the $355–$375 range.

Scenario C

The government keeps rates near the original 0.09% proposal.

Result:
Benefits are reduced further, membership declines again, and revenue remains under pressure.

Protection:
Even in this case, UnitedHealth is expected to generate around $18 billion in annual cash flow.

Stock implication:
That level of cash generation should limit downside, with the stock likely trading in the $295–$315 range rather than breaking down.

Summary

The short-term outcome depends on regulation, not demand.

The downside is supported by strong cash generation. The upside depends on how quickly the government eases pressure on the system.

The market is currently focused on the worst-case outcome. The next 12–18 months will show which path becomes reality.

Disclaimer: This is not financial advice. Investing involves risk.

The Trade Desk as a marketplace where brands buy ads across the internet.

Old world: Ford wanted TV ads → call a network → negotiate → slow, inefficient.
New world: Ford’s agency logs into TTD, uploads the ad, picks the audience (“men 25–40 who like trucks”), and presses go.

Within milliseconds, TTD’s system bids on ad slots across:

• Connected TV (Disney+, Hulu, etc.)

• Spotify, podcasts

• Websites (NYT, ESPN, etc.)

What they don’t do: they don’t buy ads inside Google Search or Meta (Facebook/Instagram). Those are closed ecosystems where outsiders can’t bid freely.

They take a small fee on every ad dollar that flows through their platform.

The moat

Google both owns a lot of the ad inventory (like YouTube) and runs one of the tools used to buy ads. That’s like the referee also owning one of the teams. Even if they try to be fair, the incentives are messy.

TTD is different:

Disclaimer: This article is for informational purposes only. It is not financial advice. I am not a financial advisor. I may buy or sell these stocks at any time. You must do your own research before investing

By reading this email, you agree to the full disclaimer found here: https://www.thevaluethesis.com/p/disclamier

• They own zero media

• They don’t own TV channels, websites, or ad space

• They’re just the buying tool

So big brands trust them more to spend money where it actually works, not where the platform owner wants it to go.

They’ve kept over 95% of customers every year for 11 years. Once agencies train on TTD, they rarely switch.

The “real” profits (GAAP)

The stock compensation

TTD pays employees a lot in stock. That cost shows up in GAAP earnings, and it matters because it dilutes shareholders over time.

In the first 9 months of 2025:

• Stock paid to employees: $378m

• GAAP profit: $256m

• “Adjusted” profit (Wall Street’s version): $589m

Why growth is slowing

Growth used to be 40%+. Now it’s more like low-20s.

Main reasons:

• Bigger base: it’s harder to grow 40% when you’re already doing $2.5B revenue.

• Market maturing: digital advertising isn’t “new” anymore.

• Management’s Q4 2025 guide implies roughly 18–20% growth (excluding political ads), which is likely the new normal.

1) It’s harder at this size
40% growth on $500m is easy math. 40% on $2.5b means finding about $1b of new revenue every year. That’s a huge ask, even for a best-in-class platform.

2) The market is maturing
Digital ads aren’t a fresh “land grab” anymore. Growth now comes more from taking share (from other platforms or old TV budgets) than from a wave of new adoption.

3) Amazon changed the game in CTV
Amazon is moving aggressively into premium streaming inventory and often makes buyers use Amazon’s own tools to access it. That locks TTD out of part of the best inventory, which limits how fast budgets can shift onto TTD.

The long-term growth range: 15% to 20%

If you’re underwriting the next 5 years, don’t anchor to the 40% era. This looks more like a steady “compounder” profile.

• 2026: roughly 16%–20% revenue growth

• 2027–2030: a gradual fade toward 14% as the company gets larger.

What keeps growth above 15%?

Management’s main levers are:

1) International
They’re still heavily US-weighted. If they truly crack Europe and Asia, that’s meaningful runway.

2) Retail media / shopper marketing
Partnering with retailers and using purchase data to target and measure results is a big growth pocket, because it ties ads to actual sales.

3) Connected TV
As cable declines, ad budgets keep moving to streaming. Programmatic buying is a natural fit here, and TTD is one of the key pipes.

What holds it back?

1) US saturation
They’re already big in the US. It’s hard to double when you’re already in almost every major agency relationship.

2) The Amazon wall
If Amazon keeps premium inventory increasingly “closed,” that’s a real ceiling on how much of the best CTV spend TTD can touch.

Valuation: where to feel comfortable

Once you accept 15%–20% as the likely long-run growth range, the price you pay matters a lot.

Scenario A: $35/share
You’re paying roughly 40x “real” earnings for 15–20% growth.

Scenario B: $25/share (your $12B–$13B market cap zone)
You’re closer to 27x real earnings.

Bottom line

The hyper-growth phase is likely over. This is becoming a high-quality compounder.
Your $12B–$13B market cap target is a sensible entry point for this phase of the business, because it matches the slower (but still strong) growth reality and gives you margin of safety.

5) My comfort-zone price

At about $35/share:

• Market cap $17.2B

• Real P/E 39x (not cheap)

At my target $12–13B market cap:

• Share price roughly $24.50–$25

• Real P/E around 27x

27x GAAP earnings for a company growing 20% with a real moat and 95% retention is reasonable.
At 39x, you’re paying extra for perfection, basically betting growth re-accelerates (AI, international, etc.).

My take

The good: best product in its niche, trusted “neutral” platform, insanely sticky customers.
The bad: heavy stock pay means your real earnings are lower and you get diluted. Growth has normalised to 20%.
I want margin of safety, wait for $25. At $35, it’s a great company, but the price is a bit stretched.

Disclaimer: This article is for informational purposes only. It is not financial advice. I am not a financial advisor. I may buy or sell these stocks at any time. You must do your own research before investing

By reading this email, you agree to the full disclaimer found here: https://www.thevaluethesis.com/p/disclamier

As I write this, markets are taking the headlines badly.

But this is exactly when you pay attention. The best chances are rare. They don’t show up often. When they do, you need to be ready.

Charlie Munger said opportunities come to a prepared mind. That’s the point: you do the work early.

Investing is mostly study, not action. Most of the time, the right move is to do nothing.

Right now, many good companies still look expensive. The prices leave no room for mistakes. So I’m building my watchlist, deciding what I’m willing to pay, and then waiting.

In late 2025, I shared Watchlist #1.

Here are 5 more businesses I’m watching, and the prices I’m waiting for in 2026

What they all have in common:

1. High Switching Costs — once you’re in, it’s hard to leave. These businesses become part of how customers operate.

2. Disconnected Narratives — the story around them is worse than the reality. Headlines and sentiment are driving the price more than the actual business performance.

3. Real Free Cash Flow — they generate cash after paying for what the business needs. That gives them options: reinvest, buy back shares, or strengthen the balance sheet.

4. A Moat I Trust — I believe they have real protection around the business. Something that makes it hard for competitors to take customers and profits.

All five are currently hated or ignored for reasons that are likely temporary or exaggerated.

And because sentiment has already pushed them closer to what I think is fair value, I’m watching these names more closely than the rest, to see if the price drops to where I’d actually want to buy.

My prices might surprise you. But my job isn’t to be optimistic. It’s to avoid big mistakes. I want a price that gives me breathing room.

That’s what a margin of safety is.

Disclaimer: This article is for informational purposes only. It is not financial advice. I am not a financial advisor. I may buy or sell these stocks at any time. You must do your own research before investing

By reading this email, you agree to the full disclaimer found here: https://www.thevaluethesis.com/p/disclamier

1. HubSpot (HUBS)

Current Price: $311 Market Cap: $16.3B

The business is solid. They generate roughly $3 billion in revenue and are free cash flow positive. Growth is steady at around 15%.

The Problem: The stock price is priced for 25%+ growth forever. At $329, the math doesn’t work. If AI slows their seat expansion or margins compress even slightly, this stock has a long way to fall.

The Price I’m Waiting For: $215 – $235 I want a margin of safety. If the price drops to this range, the risk/reward flips in my favor. Until then, I’m not interested.

2. Adobe (ADBE)

Current Price: ~$296 Market Cap: ~$124B

Adobe stock is “dead money” right now. It’s trading near $296 because the market is convinced AI will kill their business.

I disagree. Corporate workflows are built on Adobe. Switching costs are high, and compliance is messy. The financials tell a different story than the narrative: Adobe generates $23.8 billion in revenue and a massive $8.9 billion in Free Cash Flow.

The Price I’m Waiting For: $243 I am looking for a further 18-20% discount. At $243, you pay just 10x Free Cash Flow. That prices the company for zero growth. If they grow at all, you win.

3. Salesforce (CRM)

Current Price: ~$227 Market Cap: ~$212B

Salesforce is down significantly, trading around $227. The business is a cash machine, pulling in $40 billion in revenue and over $12 billion in Free Cash Flow. But even at this lower price, it’s not a screamimg buy yet.

The Price I’m Waiting For: $175 At $175, you would be buying a software monopoly for roughly 14x Free Cash Flow. That is a bargain for a company with this level of customer retention.

4. SAP (SAP)

Current Price: ~$233 Market Cap: ~$287B

SAP has a premium valuation because investors treat its cloud backlog as a safety net. Revenue is steady at $40 billion, but you are paying nearly 33x Free Cash Flow for that safety.

The Price I’m Waiting For: $116 To get interested, I need the price to be cut in half. At $116, you pay roughly 16x FCF. This treats SAP like the utility company it is, rather than a high-growth cloud stock. I want to pay for the cash, not the hype.

5. Uber (UBER)

Current Price: ~$85 Market Cap: ~$176B

Uber has successfully pivoted. They are generating $49.6 billion in revenue and $8.7 billion in Free Cash Flow. The market currently pays about 20x FCF for this.

The Price I’m Waiting For: $55 I want to pay 12–14x Free Cash Flow. At $55, you are buying a global logistics network for the same multiple as a boring industrial stock. If the economy slows, this entry price protects you.

Disclaimer: This article is for informational purposes only. It is not financial advice. I am not a financial advisor. I may buy or sell these stocks at any time. You must do your own research before investing

By reading this email, you agree to the full disclaimer found here: https://www.thevaluethesis.com/p/disclamier

The economy is strong. Ireland has a high concentration of multinational activity in tech, pharma, medical devices, and aircraft leasing, supported by a young workforce and strong export growth. Employment levels are high, wages have grown, and net migration continues to add demand.

Housing supply remains tight.

Any decent rental listing is gone instantly. Same with buying. Supply is not keeping up, and everyone feels it.

That’s the backdrop for $IRES, the largest residential landlord / residential platform in Ireland, with 3,652 homes, mostly in Dublin.

Occupancy is basically full: 99.5%.

It’s a simple model: own apartments in the most supply-constrained part of the country and collect rent.

The business has

• 3,652 units (mostly Dublin)

• 99.5% occupancy

• 99%+ rent collection

• Average monthly rent around €1,823 across the portfolio

• Net rental income margin: 78% (up 150bps)

The stock is down from €1.73 at the 2022 peak to around €0.93 today, as we head into the end of December 2025.

1) Interest rates repriced property values

When rates rise, property yields expand, and values fall. IRES even shows that Dublin PRS yields expanded 115 bps since Dec 2021, implying roughly a 32% decline in capital values.

So even if rents stay strong, your asset value can get marked down.

2) Rent controls and regulatory uncertainty

Ireland’s rent control framework has been a big overhang. IRES directly says restrictive rent regulations disproportionately impacted valuations.

You can feel this on the ground too:

• Private landlords are leaving.

• Costs are up.

• And the market assumes rental regulation will remain restrictive.

What’s changed

Two things are starting to tilt the setup.

1) Rate cuts = less pressure on valuations

IRES notes that the ECB has already cut rates multiple times since peak inflation (they reference eight cuts between June 2024 and June 2025).

Lower rates help in two ways:

• cheaper funding over time

• less pressure on property yields / capital values

  

2) The regulatory tone is shifting

In June 2025, the government announced reforms aimed at more “clarity and certainty”.

And the big practical point is this: from 1 March 2026, rents can reset to market for new tenancies (in the situations covered).

That matters because IRES believes the portfolio is under-rented vs market, they show 19% reversion.

Rent gap math

IRES currently generates approximately €86.6m in annualised rent.

Management estimates around 19% rent reversion across the portfolio as rents reset to market levels over time. This implies a potential rent gap of roughly €19–20m once turnover and resets occur.

That gives:

• Current annual rent: €86.6m

• Estimated rent reversion: €19.9m

• Implied annual rent: €103–106m

Assuming no additional asset sales, this represents the rental income capacity of the existing portfolio.

Against the current equity market capitalisation of €488m, this implies a gross rental yield in the low-20% range, before operating costs, interest, or reinvestment.

The Investment Thesis

When inflation reached close to 9% and maintenance costs increased sharply, IRES could only increase rents by about 2%. In real terms, earnings declined.

I generally prefer businesses with pricing power against inflation and competition. IRES does not fully meet that criterion, it operates within a regulated framework.

That said, it owns real assets in established Dublin locations, maintains 99.5% occupancy, and generates predictable rental income. Demand for rental housing remains high, and supply remains limited.

Disclaimer: This article is for informational purposes only and does not constitute financial advice. The views expressed here are my own and should not be taken as a recommendation to buy or sell any security. I do not currently hold a position in the stock mentioned in this report. Please conduct your own due diligence or consult a certified financial advisor before making any investment decisions.

Read more

• Price/Book Ratio = 0.74x.

• P/E Ratio (TTM) = 7.5x

• 3.7% dividend yield.

Total assets sit at $345.7 million. While the bank slightly shrank overall, the loan book increased to $275.5 million, and deposits rose to $280.7 million. Crucially, they lowered their interest expense, meaning they paid less to fund the bank than they did last year. With total equity hitting $40.7 million, the book value is now $40.64 per share. Directors have over $27 million of their own cash sitting in the bank’s deposits.

Read more

Gartner ticks all those boxes. a strong competitive advantage, excellent cash flow, and a price that offers a margin of safety.

Here is a look at how the business actually makes money, and why I think it is currently selling for less than it is worth

1. The Business Model: High Margins and Upfront Cash

Gartner sells research and advice to corporate executives.

• Operating Leverage: They produce intellectual property (reports and data). Once a report is written, the cost to sell it to one client or one million clients is the same. This allows for very high profit margins on every new subscriber.

• Recurring Revenue: Most of their revenue comes from multi-year subscriptions, making the business predictable and stable.

• Negative Working Capital: Clients pay their bills upfront at the start of the contract. This means Gartner collects the cash before they deliver the service, giving them a constant “float” of money to invest or return to shareholders.

  

2. The Competitive Advantage

Their primary product (the “Magic Quadrant”) has become the industry standard for rating technology vendors.

• For Buyers: Corporate executives rely on these ratings to justify billion-dollar spending decisions to their boards. It provides them with necessary independent verification.

• For Sellers: Technology companies effectively must subscribe to Gartner to understand how to be rated. If they are not rated, they risk being invisible to buyers. This creates a self-reinforcing loop that is very difficult for competitors to break.

Imagine you’re the head of technology at a big company, and you’re facing a high-stakes, $50 million decision on what new software to buy

• if you pick the wrong software, the company loses money, and your job is at risk.

• If you pick the right software, you just did your job.

So, what do you do? You pay Gartner a subscription fee to see their report cards (the famous “Magic Quadrant”).

If Gartner says a software is a leader, and you buy it, you are safe. Even if the software fails, you can tell your boss, “I bought what the experts recommended.”

That is why managers are happy to pay Gartner’s fee: It is a small price to pay for peace of mind.

It is essentially “Career Insurance” for executives.

In summary:

• It’s a Subscription: They collect money every year, recurring like a utility bill.

• It’s Highly Profitable: They write a report once and sell that same PDF to 15,000 different clients.

• Cash Upfront: Customers pay the bill at the start of the year (negative working capital).
  

3. Why Is It On Sale?

The stock is currently trading at a discount because of two market fears:

• Cyclical Headwinds: Gartner charges per user. As the tech sector has laid off employees recently, the number of paid subscriptions has temporarily dropped.

• Structural Fear (AI): Investors worry that Artificial Intelligence will replace the need for human analysts, destroying Gartner’s pricing power.

Disclaimer: The author holds a long position in Gartner (IT). This content is for informational purposes only and should not be considered professional financial advice. All investments involve risk.

4. Why I Think The Market Is Wrong

Read more

Right now, that comparison is getting harder. The S&P 500 is trading around 25x forward earnings, well above its historical average. Great companies at a fair price are almost nowhere to be found.

When the market gets expensive, I continue to study companies and get my watchlist ready. I demand a high margin of safety, so these stocks might never reach my price. But if they do, I would consider adding them.

Read more

But if you zoom in to October 2025, the chart tells a different story. The stock experienced a significant contraction, shedding nearly 40% of its value following a Q3 earnings call where the new leadership team reset expectations.

The market has effectively repriced Fiserv from a “Growth Compounder” to a “Distressed Utility.” This analysis explores the mechanical reasons behind this drop, the structure of the business, and the valuation metrics that currently define the stock.

1. The Business: Two Engines

According to recent 10-K filings, Fiserv’s ~$19–$20 billion in revenue is derived essentially from two distinct business models.

Segment A: Financial Solutions

• Approx. Revenue Split: 47%

• The Function: This segment acts as the operating system for financial institutions. Products like DNA, Premier, and Signature serve as the “Core Banking” platforms. When a consumer checks their balance or transfers funds, Fiserv’s software often manages the data.

• The Structural Dynamic: High switching costs. Migrating a bank’s core ledger is operationally complex, expensive, and risky for a bank. Consequently, client retention rates in this segment have historically hovered near 99%.

Segment B: Merchant Solutions

• Approx. Revenue Split: 53%

• The Function: This segment handles payment processing for businesses.

  • Small Business: The Clover point-of-sale system (hardware + software).

  • Enterprise: The Carat platform, which manages payments for large global merchants.

• The Structural Dynamic: Revenue is driven by transaction volume and software subscriptions. Distribution is heavily aided by bank partnerships (e.g., a bank offering Clover to its business clients), providing a distinct channel compared to digital-first competitors.

2. What Happened in October?

For several years, the market priced Fiserv based on expectations of consistent “double-digit” earnings growth. To achieve these targets, the company focused heavily on margin expansion.

In October 2025, the new CEO, Mike Lyons, provided a candid assessment of the company’s infrastructure. The core message was that maintaining long-term competitiveness requires significant immediate investment.

• The Pivot: Management announced a need to increase capital expenditures to modernize legacy banking platforms.

• The Consequence: This spending is expected to compress margins in the near term. The company labeled 2026 a “transition year,” leading investors to revise their earnings models downward.

3. The Valuation Data & Expected Return

The stock is currently priced as if growth will permanently stagnate.

Valuation Snapshot (Late 2025 Context)

• Current Stock Price: $63–$65

• Forward Earnings (2025 Guidance): $8.55/share (adjusted EPS)

• P/E Ratio: 9.5x (Historic Low)

  • S&P 500 Average: 24x

  • Visa/Mastercard: 30x+

  • Fiserv 5-Year Average: 17x

Read more

Price-to-Earnings: 7.8×

If you are looking for the next tech rocket ship, this one isn’t for you.
Sirius XM is a mature, slow-moving monopoly on satellite radio

The thesis here isn’t about growth, it’s about capital return. You’re buying a business that generates more than a billion dollars in free cash flow every year and sends most of it back to shareholders through dividends and buybacks.

The Business

Sirius XM is really two businesses masked as one:

• The Satellite Business: This is the core. It puts radio in cars. It has high margins (59% gross margin ) and incredibly sticky customers (churn is only 1.6% ). Even though everyone has Spotify on their phone, 33 million people still pay for Sirius because it’s easy, curated, and integrated into the dashboard.

• Pandora & Off-Platform: This segment fights the brutal streaming wars against Spotify and Apple. It operates at much lower margins (31% gross margin ). The bright spot? While music streaming revenue is dipping, their Podcast business is booming, growing enough to actually pull total segment revenue up 1% this quarter.

Free Cash Flow

If you check the portfolios of top value investors, you will spot Sirius XM. I think the reason is the Free Cash Flow. Despite revenue declining slightly (-1% YoY), the company is exceptionally converting those revenues into free cash flow.

• Q3 Free Cash Flow: $257 Million.

• 2025 Guidance: They just raised it to $1.2 billion.

Because the satellite network is already built and the major costs are behind them, they don’t need heavy reinvestment to keep the service running. Most of the cash generated each quarter can flow straight through to shareholders instead of being recycled back into the business.

Dividends + Buybacks

The Dividend: At current prices, the stock yields about 5%.

The Buybacks: Second, share repurchases have resumed. After pausing during the Liberty Media split, they bought back about 860,000 shares in Q3 2025, totaling roughly $20 million

Even if the business shrinks 1% a year, as long as they reduce the share count by 4–5% a year, your personal slice of the profit actually gets bigger.

Risks

• The “CarPlay” Threat: While Sirius has a monopoly on satellite radio, they face “substantial competition” from tech giants (Amazon, Apple, Google). Every car with Apple CarPlay makes it easier to cancel Sirius for Spotify.

• Auto Sales Dependence: Sirius depends on new car sales to acquire subscribers. If the auto industry slows down—whether due to tariffs or broader economic issue, the number of people starting free trials will decline.

• Debt: They carry significant debt ($9.0B ), with a leverage ratio of 3.8x. In a high-rate environment, that debt gets more expensive to service.

• Satellite Risk: While CapEx is down, it isn’t zero. In the first nine months of 2025, they spent $186 million just on satellite construction. If a satellite fails (like SXM-7 did), it is a multi-year, capital-intensive problem to fix.

Bear Case

The bear argument rests on the view that the business is in a slow, irreversible decline due to technological shifts and changing consumer habits.

• Revenue: The financials confirm a slow contraction. Total revenue dropped 1% in Q3 2025 to $2.16 billion, and is down 2% year-to-date.

• Subscriber Loss: The core engine is losing paying customers. SiriusXM self-pay subscribers decreased by 40,000 in Q3 2025. Pandora is seeing steeper declines, with Monthly Active Users dropping 5% year-over-year to 41.6 million.

• Demographic & Tech Headwinds: The company explicitly acknowledges that “changing consumer behavior” and “new technologies” (like in-dash infotainment) are risks that could reduce subscribers. Substantial competition from streaming giants like Spotify, Apple Music, and YouTube, which offer free or low-cost alternatives that appeal to younger demographics.

Bull Case

The bull argument relies on the company’s exceptional ability to generate cash and return it to shareholders.

• Extremely Low Churn: Despite the competition, SiriusXM’s self-pay monthly churn remains remarkably low at 1.6%. This indicates a very “sticky” core customer base that is not easily defecting to streaming.

• Massive Free Cash Flow: The company is a cash generator. It produced $257 million in free cash flow in Q3 alone and raised its full-year 2025 guidance to approximately $1.225 billion.

• Capital Returns (Dividend + Buybacks): Management is using this cash to pay shareholders.

  • Dividends: They declared a dividend of $0.27 per share in Q3 2025.

  • Buybacks: After the Liberty Media transaction, repurchases resumed. In Q3 2025, the company bought back 860,448 shares for approximately $20 million.

Right now, shareholders are getting roughly a 6% annual return, mostly from the quarterly dividend of $0.27 per share and the buybacks that restarted in Q3, when the company retired about $20 million worth of stock.

Disclaimer: This article is for informational purposes only and does not constitute financial, investment, or legal advice. Analysis is based on public SEC filings and earnings reports; however, accuracy is not guaranteed and market conditions change. Investments involve significant risk, including the potential loss of principal. Always conduct your own due diligence and consult a qualified financial advisor before making investment decisions. The author holds no position in the securities mentioned.

The market tends to overreact to short-term changes, and by short term, I mean a few quarters that create uncertainty. So this kind of swing doesn’t surprise me.

My job is to separate noise from fundamentals to understand what really changed, what didn’t, and whether there’s value left beneath the panic.

The recent decline comes down to slowing growth in North America, cut guidance, and rising competition. The headlines sound bad. But when you look beyond them, the business still stands strong — both financially and operationally.

Do we really understand what’s happening here — and why the stock is down?

When I say a stock is down, I’m really talking about the market’s perception of the company’s value, not just a price move on a screen.

But what actually makes a company’s value decline? It happens when earnings power weakens, the balance sheet deteriorates, the long-term outlook becomes riskier, or when the brand starts to lose relevance, innovation slows, competition tightens, and the moat erodes.

You could expand that list further, macro shifts (rates, spending cycles), operational mistakes (poor execution, inventory build-up), or capital decisions (buybacks at the wrong time, over expansion). But the core idea stays the same: the value falls only when something real and lasting changes beneath the surface.

Subscribe now

So, do I see any of that here with Lululemon? Let’s look.

What I see is a premium global brand, with massive expansion potential in Europe and China, a strong balance sheet, cash in hand, and better margins than most of its peers — around 60% gross margins, a billion dollars in cash, and zero debt. The company is buying back shares, expanding internationally, and doing it profitably.

Then what’s the problem?

U.S. growth has slowed, consumers have become more selective, and management made a few mistakes: slower product refreshes, higher inventories, and some pricing missteps. The market punished those hard.

But those are fixable issues, not structural ones. The business underneath still looks strong.

A Split Story

If you’ve listened to the company’s recent earnings calls or gone through the filings, you’ll notice something interesting, it’s really two stories in one.

International growth is running in the mid-20s, while the Americas are soft. That’s the split. So the natural question is — why is the business growing strongly overseas but slowing at home? And why is that hitting the bottom line?

I’d say it’s a mix of things — tariffs, supply chain costs, inventory build-up, and a cut in guidance.

Each one on its own wouldn’t break the story, but together they’ve pressured margins and spooked investors who were expecting smoother execution.

The Americas are weak — consumers have turned cautious, and competition in premium activewear has intensified, especially as Nike and Adidas leaned heavily on discounts this year. Consumer sentiment has shifted too. The sense of newness just wasn’t there for Lululemon, traffic held up, but visitors didn’t quite feel the excitement of fresh products.

International markets are strong, China and Europe both delivered double-digit growth, expanding quickly in regions that are still far from saturation. The brand remains unknown to many potential future loyal customers, which makes this a genuinely exciting story, if it’s executed well.

That global diversification is exactly what gives Lululemon its resilience. When one region slows, others pick up the pace.

Think of Coca-Cola, Nike, or Apple, brands that travel effortlessly across borders, cultures, and lifestyles. There’s something special about that. Most products can’t even scale smoothly between U.S. states, let alone countries.

Lululemon has that rare potential, the ability to become a truly global lifestyle brand, not just a retailer. That’s a powerful advantage over time.

The company has now reached nearly 800 stores globally, with new openings supported by solid demand and strong unit economics. This is profitable growth, not expansion for the sake of expansion.

Even as the company grows rapidly internationally, its margins remain among the best in retail. Yes, they’ve dipped slightly because of tariffs and higher costs, but at around 58%, they’re still well above peers like Nike and Adidas.

That kind of margin strength says a lot, it means the brand still commands pricing power, and more importantly, it still has the trust and loyalty of its customers.

Is the Market Right to Value Lululemon at $20 Billion?

Read more

Just picture this, all in one app:

• Payments & QR: $PYPL, $SQ, Apple Pay, Google Pay

• Banking & savings: Revolut, Chime, Monzo

• E-commerce marketplace: $AMZN, $SHOP, $EBAY

• Grocery & delivery: $DASH, $UBER, Instacart

• Travel: $BKNG, Expedia

• Classifieds & real estate: AutoTrader, $Z

• Advertising for merchants: $GOOG, $META

• SME lending & BNPL: $AFRM, Stripe Capital, $PYPL Credit

• Government & utilities: driver’s licence, tax payments, electricity, water — all inside the same app

Now imagine logging into one single app, one ID, one wallet where you can pay your bills, book a trip, advertise your store, buy groceries, or apply for a small business loan, all without leaving the platform.

Add AI on top of it.
It knows your full transaction history, spots fraud instantly, manages your cash flow, underwrites credit, and suggests better financial moves. It becomes your operating system for daily life.

That’s exactly what Kaspi ($KSPI) has already built in Kazakhstan, while most Western markets are still trying to stitch together a dozen different apps to do the same thing.

Investment Thesis

Kaspi’s biggest strength is its three-sided network effect, connecting consumers, merchants, and payments under one roof, with Kaspi Pay at the core.

The more people use the app, the more merchants join the marketplace.
As more merchants accept Kaspi Pay, the app becomes more useful for consumers.
Each side strengthens the other and that flywheel creates a strong moat.

Merchants can grow their business directly through Kaspi, take loans, advertise, and expand their reach. Consumers can shop, pay, and get instant financing. Every part of the ecosystem talks to each other, creating a level of integration that’s rare even by global standards.

I see limited downside here, backed by predictable earnings, strong free cash flow, and zero need for outside capital.
Even after the Hepsiburada $HEPS acquisition, Kaspi still sits on a net cash position.

At today’s valuation, you’re looking at a cash-rich, dominant platform trading at single-digit earnings multiples

Business Overview

Kaspi.kz runs two main apps:

• Kaspi Super App (Consumer) — used by millions daily for payments, shopping, travel, transport, and finance.

• Kaspi Pay (Merchant) — built for entrepreneurs and SMEs, offering QR payments, delivery, accounting, loans, and advertising.

Engagement is through the roof: on average, users make 75 transactions per month inside the app — that’s how deeply embedded it is in daily life.

Kaspi now owns 66.35% of Hepsiburada in Türkiye and is acquiring Rabobank A.Ş, which gives them a local banking licence. Türkiye is early-stage, but if they replicate even half of what they’ve done in Kazakhstan, it’s meaningful upside.

1H 2025: Solid Growth, Despite Headwinds

Headline results (excluding Türkiye):

• 2Q 2025 revenue: KZT 734B (+20% YoY)

• 2Q 2025 net income: KZT 276B (+14% YoY)

• 1H 2025 revenue: KZT 1.4T (+20% YoY)

• 1H 2025 net income: KZT 535B (+15% YoY)

Including Türkiye:

• 1H 2025 revenue: ≈ KZT 1.8T ($3.5B)

• 1H 2025 net income: ≈ KZT 513B ($987M)

By segment:

• Payments: TPV +21% YoY; revenue +16%; net income +19%.

• Marketplace: GMV +15%; revenue +25%; take rate 10%. e-Grocery up 57% YoY. Smartphones weaker due to IMEI registration law, but excluding smartphones, e-commerce GMV grew 31% YoY showing strong core demand.

• Fintech: TFV +17%; revenue +21%; net income +8%. Deposit rates rose 26% YoY

Capital Allocation & Türkiye

2025 is an investment-heavy year:

• The final $526.9M payment for Hepsiburada closed in July. ( $HEPS )

• The Rabobank A.Ş acquisition in Türkiye is expected to close in 2H 2025, with roughly $300M in capital deployed.

From 2026 onward, Kaspi plans to balance growth investments with capital returns, and I expect the company to start paying dividends to shareholders.

Valuation — Growth

Kaspi earned around $2.0B in 2024.
Assuming 15% net income growth, that puts:

• 2025E net income: $2.3B

• 2026E net income: $2.65B

• Shares outstanding: 190M

At a $15B market cap ($79/share):

• P/E 2025E: 6.5×

• P/E 2026E: 5.7×

• Net cash: $2B +

You’re paying mid-single-digit earnings multiples for a super app that’s still growing mid-teens, with a clean balance sheet.

On a free cash flow basis, adjusted for loan outflows, Kaspi’s core FCF is around $2.9B, which puts it at roughly 5× core FCF.

Even using conservative assumptions, growth slowing from mid-teens to high single digits by the end of the decade the stock looks deeply undervalued.

You’re basically paying value-stock multiples for a business with growth, durability, and compounding power.

Risks

I think the biggest risk here is geography. Kazakhstan sits close to Russia, and that proximity will always keep a small discount on the stock — it’s just part of investing in the region.

Regulation and deposit costs are the other two things to watch. The new 10% tax on government bond income and higher reserve requirements already shaved about 200 basis points off margins this year. Rising rates also pushed deposit costs up roughly 26% year over year. It’s not alarming, but it’s worth keeping an eye on if the rate environment stays elevated longer than expected.

As for competition, I’m not worried. If you look at Kaspi’s daily and monthly user data, they’re far ahead of Halyk Bank or anyone else in digital engagement.

The other area that needs time is execution in Türkiye. Hepsiburada is improving, but integrating the banking operations through Rabobank A.Ş will take work. It may weigh on margins in the short term, but the long-term opportunity looks meaningful.

Takeaway

Kaspi is one of the few true super apps in the world a fintech, marketplace, and bank, all rolled into one.

You’ve got:

• Proven management with a compounding mindset.

• A three-sided platform built around real user needs.

• Predictable earnings and strong free cash flow.

• A fortress balance sheet with net cash.

• Insider ownership and minimal stock-based comp.
  
  
  

  ---

  
  Disclaimer

  The author of this analysis is a current shareholder in Kaspi.kz ($KSPI). The views and opinions expressed here are personal and for informational purposes only. This does not constitute financial or investment advice. All investment decisions should be made based on your own research and due diligence.

Read more

• 2023: +33%

• 2024: +36%

• 2025 YTD (as of November 1): +32%

• Since 2023: +139% vs. S&P 500 +61%

Read more

A Business Built for Efficiency

Molina runs managed-care health plans for people covered under government programs, mainly Medicaid and Medicare. It also participates in state insurance marketplaces under the Affordable Care Act.

Roughly three-quarters of its revenue comes from Medicaid, where the company works with state governments to deliver care for low-income families.
This business isn’t flashy, but it’s steady.
Molina’s moat lies in its deep operational expertise, understanding complex state rules, keeping administrative costs low, and running an efficient network while maintaining quality of care.

For years, that efficiency has been its competitive edge.

What Went Wrong This Year

After several strong years of growth, 2025 brought a sharp increase in healthcare costs.
People who delayed medical treatments during the pandemic returned for care, many of them sicker and requiring more expensive procedures. Specialty drug costs also rose sharply, while hospitals passed along higher wages and inflation-driven expenses.

All this hit Molina’s profitability at once.
The medical cost ratio, essentially the share of every premium dollar spent on care, rose above 91%, up from the company’s usual range of 88–90%.
That small change in percentage terms has a large impact on margins.

The ACA (Marketplace) business was hit hardest. Medical costs there rose faster than premiums, forcing Molina to cut its footprint in weaker markets and raise prices significantly to restore profitability.

Why It’s Likely Temporary

Healthcare inflation isn’t permanent.
State reimbursement rates are already being reset for 2026 with more current cost data. Molina is exiting the least profitable counties and repricing others. Medicaid, which drives most of earnings, remains solidly profitable.

As demand normalizes and state rate adjustments catch up, the pressure on margins should ease.
This isn’t a structural problem, it’s timing.

What Management Is Doing Right

Rather than chase growth, Molina is staying disciplined. It’s shrinking exposure in unprofitable ACA markets, raising premiums where necessary, and continuing share repurchases while the stock trades well below intrinsic value.

That’s a rational use of capital, buying more of a business that remains strong at a time when sentiment is weak.

The balance sheet is solid, giving management room to invest and stay patient.

Investment Thesis

Molina’s stock has fallen about 46% in 2025. The market seems to think the business is in trouble. I don’t completely disagree, but I believe the issue is temporary. To me, it looks more like the company is dealing with a short-term gap between rising medical costs and slower state rate adjustments.

This post reflects my own opinions and research. It is shared for educational purposes only and does not constitute investment advice. Investors should always conduct their own due diligence or consult a professional adviser before making decisions.

Read more