Bill Ackman vs. Chris Hohn: Inside the $10B Microsoft Stock Trade

In the same quarter, Bill Ackman built a $2.1 billion Microsoft position, and Chris Hohn’s TCI sold $8 billion. Same business. Opposite trades. The fact pattern that decides which one is right is sitting in plain sight in Microsoft’s filings. Here is the math, the misread, and the framework you can copy.

TL;DR

• Two billionaires made opposite trades on Microsoft in Q1 2026: Pershing Square bought $2.1B; Chris Hohn’s TCI sold roughly $8B. Same name, same quarter, opposite thesis.
• Ackman paid about 21x forward earnings — the market multiple and well below Microsoft’s 5-year average of 32x. Subtracting the ~$200B OpenAI stake from the market cap implies an effective operating business multiple closer to 19.5x.
• Hohn sold because he fears AI will disrupt the Office franchise; Ackman bought because he reads M365 as a sticky procurement default. The IT-procurement evidence currently favors Ackman.
• Microsoft is now Pershing’s #4 holding at 15.3% of the disclosed equity book, behind Brookfield (17.6%), Amazon (17.4%), and Uber (15.7%).
• The trade extends a three-year pattern — Alphabet 2022, Amazon 2025, Meta 2025, Microsoft 2026 — each one bought into mechanical selling on a quality compounder.
• Read the contracted backlog honestly: $627B in commercial RPO grew 99% headline, but excluding OpenAI commitments grew only ~26%, and the sequential change was only $2B.

The setup: A $2.9 trillion company at the market multiple

Two numbers carry the entire trade. Microsoft sold off after its fiscal Q2 2026 print on January 28. Pershing Square began building its position in February at roughly 21 times forward earnings — the same multiple as the S&P 500. Microsoft’s own 5-year average sits closer to 32 times.

Run the subtraction of the headline P/E skips. Microsoft’s market cap is roughly $2.9 trillion. Ackman values the company’s ~27% economic interest in OpenAI at ~$200 billion — that asset produces zero current earnings and does not flow into any reported EPS. Strip it from market cap, and you get a $2.7 trillion operating business. Divide by the same forward earnings denominator, and the effective entry multiple drops to roughly 19.5x. The market is paying 21x for a business that, on its operating economics alone, is priced below the index.

The funding source tells you the conviction level. Pershing sold roughly 95% of its Alphabet position to free capital for Microsoft. The position now sits at 15.3% of the disclosed equity book — Pershing’s #4 holding, behind Brookfield, Amazon, and Uber. That is not a starter position. That is a high-conviction reweighting from one mega-cap quality compounder to another.

Two billionaires, one stock, opposite trades

Pershing Square’s $2.1 billion buy is only half the story. In the same quarter, Sir Chris Hohn’s TCI sold roughly $8 billion of Microsoft, cutting the position from ~10% of the fund to under 1%. Two of the most respected concentrated long-only managers in the world, opposite sides of the same stock, same quarter. The divergence is not noise. It is the trade.

Hohn’s thesis: AI-driven workflow tools will erode the Office franchise. TCI flagged Office productivity as the structural risk, arguing that new AI-native platforms could reroute knowledge-work spend away from M365. TCI rotated proceeds toward Alphabet.

Ackman’s thesis is the exact opposite. The bundle is sticky, the procurement friction is enormous, and the AI overlay is a pricing leg — not a disruption vector. Microsoft disclosed that more than 120,000 custom Copilot agents had been deployed across enterprises by Q1 2026. Each custom agent embeds Copilot deeper into a customer’s workflow, identity model, and compliance stack. The displacement cost rises with every deployment.

The resolution is empirical, not philosophical. If knowledge work spends rotate to AI-native tools outside the Microsoft stack within the next 24 months, Hohn is right. If 120,000 custom agents compound to 500,000 and Copilot seats keep climbing inside existing M365 contracts, Ackman is right. Watch the seat count and the agent count. The trade settles itself.

M365: A sticky default, not just a bundle

Microsoft reported 450 million Microsoft 365 commercial seats in fiscal Q2 2026. Paid M365 Copilot adoption sits at 15 million seats — about 3.3% of the addressable base. The runway is the gap between those two numbers, and it widens further if Microsoft moves to include Copilot in base M365 enterprise licenses by 2027 — a path that several analysts now treat as the baseline assumption.

Pricing tells the same story. M365 Business Standard runs $12.50 per user per month. Business Premium runs $22. Copilot for enterprise is a $ 30-per-user-per-month add-on. Buying Word, Excel, PowerPoint, Outlook, and Teams from five separate vendors is not a credible procurement path in any Fortune 500 IT shop. The bundle is cheap relative to substitutes, sticky relative to switching costs, and embedded in compliance frameworks.

Copilot does not need to be the best AI model in the world. It needs to be the default that an IT director never has to defend to compliance. That low-friction default is the pricing leg the bear case ignores — and the mechanism Hohn is betting against.

Azure: A model-agnostic toll road, with a backlog you have to read honestly

The market read Microsoft’s restructuring of OpenAI as a concession. Ackman reads it as a deliberate pivot to multi-model architecture. The data support him. Microsoft disclosed that more than 10,000 enterprise customers now run multiple models on Azure Foundry, and the platform catalog spans more than 1,900 models — foundation, reasoning, small language, multimodal, and industry-specific.

Azure grew 39% reported and 38% constant currency in fiscal Q2 2026. Microsoft AI revenue is at a $37 billion annual run rate, up 123% year over year. That is the franchise the market is supposedly worried about.

Read the contracted backlog honestly

The bull-case talking point is $627 billion of commercial remaining performance obligations, up 99% year over year. The honest read needs two adjustments.

First, excluding OpenAI-related commitments, RPO grew approximately 26% year over year — much closer to the historical enterprise-cloud contract trajectory. A material slice of the headline backlog is Microsoft’s customer (OpenAI) committing to spend with Microsoft. That is real revenue, but the circularity matters.

Second, the sequential change was small. Three months earlier, in fiscal Q2, the backlog stood at $625 billion. The Q3 number is $627 billion — a $2 billion sequential add. The bookings deceleration that nobody is plotting against the headline.

Neither adjustment kills the thesis. Both belong in any honest write-up of it.

The $190 billion CapEx number

This is the slide that scared the Street. Microsoft guided to roughly $190 billion of capital expenditure for calendar 2026 — a 61% step-up from 2025, including roughly $25 billion attributable to higher memory and component pricing. The headline is shocking. The breakdown is not. Roughly two-thirds is growth CapEx — servers and networking equipment with a direct revenue tie. The remainder is broader infrastructure and facilities.

The contemporaneous criticism of early AWS and Google Cloud build-outs was identical: too expensive, too risky, too soon. Ten years of hindsight reframed both as a direct investment in infrastructure dominance. Ackman is making the same bet. Hohn is calling it a top-tick. The CapEx schedule does not resolve that argument. The next four quarters of Azure-ex-OpenAI growth will.

Why quality mega-caps keep getting mispriced

Two forces converge on every one of these dislocations.

The structural force is index ownership. Every dollar invested in an S&P 500 fund must buy or sell Microsoft, regardless of fundamentals. The reactive force is volatility-intolerant capital — multi-strategy pod shops cutting risk, quants chasing momentum, and discretionary funds with quarterly drawdown limits.

Both forces meet at the same trigger. A soft quarter or a CapEx scare sets off mechanical selling. The business did not change. The shareholder base rotated. When mechanical flows collide with a guide-down, dominant franchises get temporarily mispriced. That is the gap. Patient owners step in. Volatility-intolerant capital steps out. The transfer of ownership is the trade.

This is not Ackman’s first rodeo — and it is the smallest of the four

The Microsoft trade fits a three-year sequence:

• Late 2022 — Alphabet. ChatGPT launches. The Street panics that Google search is dead. Pershing buys Alphabet in Q1 2023 — the position later compounded more than 300%.
• Spring 2025 — Amazon. Liberation Day tariffs hit. AWS margins get questioned. Pershing buys. The position is up roughly 36% from cost.
• Late 2025 — Meta. Management guides to bigger CapEx than the Street expected. The stock sells off. Pershing buys.
• February 2026 — Microsoft. The fiscal Q2 print spooked the market over Azure deceleration and $190B in CapEx. Pershing buys 15.3% of the equity book — the fourth-largest position, smaller than those of Brookfield, Amazon, or Uber.

Four trades. Four short-term panics. Each one ended with a dominant compounder bought at a discount to its historical multiple. Microsoft in February is this quarter’s example. The underlying playbook is what ports forward.

The three-step playbook

Step 1 — Identify before you need to act

Ackman did not screen for Microsoft in February. The thesis was ready. Track dominant compounders long before the dislocation arrives. Build the watch list. Know which ranges you will pay for and which you will not.

Step 2 — Wait for mechanical selling to do the work

You cannot manufacture the buying opportunity. Short-term holders and volatility-intolerant capital create it for you. The trigger is almost always the same shape — a soft quarter, a guidance scare, a tariff headline, an AI competitive panic — followed by index-flow and pod-shop selling. Wait for it.

Step 3 — Execute at a price that respects the next decade

Twenty-one times forward earnings on a business compounding M365 at mid-teens and Azure at near 40% is not an aggressive price. Strip out the OpenAI stake, and you are paying closer to 19.5x for the operating business. That is the discount you waited for. Pay it. Microsoft in February 2026 is a live example. The playbook itself is timeless.

How the VAP framework slots Microsoft in

Every position in the ValueAligned Portfolio fits one of three buckets:

· Ownership companies — capital-light compounders with owner-aligned management and high incremental return on capital.

· Monopolies — infrastructure choke points the world cannot route around at any reasonable cost.

· Cannibal companies — aggressive share shrinkers buying back stock at a discount to intrinsic value.

Microsoft uniquely straddles the ownership company and the monopoly. M365 is the operating layer of knowledge work; Azure is increasingly the model-agnostic toll road for enterprise AI. Rare combinations are worth waiting for at a price that respects them. The VAP has held Microsoft well before Ackman’s February entry — Ackman’s trade is corroborating evidence, not the reason for the position.

I was working with a couple last quarter who asked me whether they should chase momentum names or wait. The answer was neither. Build the list. Set the price. When the volatility-intolerant money runs, you act. That is the trade Ackman just placed. It is the same trade available to any disciplined long-term owner — provided you have done the watch-list work before the dislocation arrives.

Market noise vs. Ackman’s read (and Hohn’s counter)

Key takeaways

• Two top concentrated managers took opposite sides of Microsoft in Q1 2026 — Ackman bought $2.1B, Hohn’s TCI sold $8B — on opposite reads of the M365 franchise.
• Ackman’s 21x forward entry implies an effective operating-business multiple closer to 19.5x after subtracting the ~$200B OpenAI stake.
• 120,000+ custom Copilot agents and the procurement-friction mechanic are the empirical evidence that currently favors Ackman over Hohn — though the next four quarters of Azure-ex-OpenAI growth will resolve the argument.
• Read the $627B RPO honestly: ex-OpenAI growth was ~26% and sequential change was $2B. The bookings story is strong but more nuanced than the headline.
• The three-step playbook — identify, wait, execute — is the part that ports directly to any disciplined long-term portfolio.

Endnotes

1. “Pershing Square bought $2.1B; Chris Hohn’s TCI sold roughly $8B” — Hedgeweek coverage of TCI’s $8 billion Microsoft sale on AI-disruption concerns. https://www.hedgeweek.com/tci-slashes-8bn-microsoft-position-amid-ai-disruption-concerns/

2. “about 21x forward earnings” — Investing.com / Yahoo Finance coverage of Ackman’s Microsoft thesis at 21x forward. https://finance.yahoo.com/markets/stocks/articles/microsoft-undervalued-why-bill-ackman-121939551.html

3. “fears AI will disrupt the Office franchise” — HedgeCo Insights on TCI’s Office-disruption thesis driving the Microsoft sale. https://hedgeco.net/news/05/2026/tci-slashes-microsoft-stake-over-ai-disruption-risk.html

4. “#4 holding at 15.3% of the disclosed equity book” — ValueSider Q1 2026 13F portfolio breakdown for Pershing Square. https://valuesider.com/guru/bill-ackman-pershing-square-capital-management/portfolio

5. “Alphabet 2022, Amazon 2025, Meta 2025, Microsoft 2026” — 24/7 Wall St. recap of Ackman’s pattern of buying mega-cap quality into AI-related sell-offs. https://247wallst.com/investing/2026/05/15/bill-ackman-bought-google-after-the-chatgpt-panic-up-300-he-bought-amazon-after-liberation-day-up-36-now-hes-backing-up-the-truck-again/

6. “excluding OpenAI commitments grew only ~26%” — Motley Fool analysis isolating Microsoft’s ex-OpenAI RPO growth from the 99% headline. https://www.fool.com/investing/2026/05/03/the-biggest-reason-microsofts-earnings-report-was/

7. “began building its position in February” — Fortune coverage of Pershing Square’s February 2026 Microsoft accumulation. https://fortune.com/2026/05/15/bill-ackman-microsoft-stock-openai-azure-spending/

8. “5-year average sits closer to 32 times” — FinanceCharts historical MSFT PE data. https://www.financecharts.com/stocks/MSFT/value/pe-ratio

9. “values the company’s ~27% economic interest in OpenAI at ~$200 billion” — IndexBox analysis of Microsoft’s 27% economic interest in OpenAI as referenced in Ackman’s letter. https://www.indexbox.io/blog/microsofts-27-openai-stake-valued-at-significant-level/

10. “sold roughly 95% of its Alphabet position” — QuotedData reporting on Pershing Square’s Alphabet sale to fund the Microsoft position. https://quoteddata.com/2026/05/pershing-squares-bill-ackman-sells-top-holding-in-alphabet-to-fund-2-1bn-investment-in-microsoft/

11. “sold roughly $8 billion of Microsoft — cutting the position from ~10% of the fund to under 1%” — Hedgeweek detail on the magnitude and pacing of TCI’s Microsoft sale. https://www.hedgeweek.com/tci-slashes-8bn-microsoft-position-amid-ai-disruption-concerns/

12. “TCI flagged Office productivity as the structural risk” — HedgeCo Insights on Hohn’s specific concern that AI-native tools could erode Office. https://hedgeco.net/news/05/2026/tci-slashes-microsoft-stake-over-ai-disruption-risk.html

13. “more than 120,000 custom Copilot agents deployed across enterprises by Q1 2026” — Stackmatix 2026 Microsoft Copilot adoption statistics. https://www.stackmatix.com/blog/microsoft-copilot-adoption-statistics-2026

14. “450 million Microsoft 365 commercial seats” — Office365 IT Pros analysis of Microsoft’s fiscal Q2 2026 results. https://office365itpros.com/2026/01/30/microsoft-fy26-q2-results/

15. “15 million seats — about 3.3% of the addressable base” — Motley Fool reporting on Microsoft’s fiscal Q2 2026 Copilot adoption disclosure. https://www.fool.com/investing/2026/02/09/microsoft-finally-revealed-how-many-paying-copilot/

16. “include Copilot in base M365 enterprise licenses by 2027” — Stackmatix analyst projection for Copilot bundling into base M365 enterprise plans. https://www.stackmatix.com/blog/microsoft-copilot-adoption-statistics-2026

17. “Copilot for enterprise is a $30 per user per month add-on” — Microsoft 365 Copilot enterprise pricing page. https://www.microsoft.com/en-us/microsoft-365-copilot/pricing/enterprise

18. “More than 10,000 enterprise customers now run multiple models on Azure Foundry” — CIO Dive coverage of Microsoft’s fiscal Q3 2026 cloud and AI metrics. https://www.ciodive.com/news/microsoft-AI-cloud-growth-earnings-Q1/804338/

19. “more than 1,900 models” — Microsoft Learn — Azure AI Foundry Models overview. https://learn.microsoft.com/en-us/azure/foundry/concepts/foundry-models-overview

20. “39% reported and 38% constant currency in fiscal Q2 2026” — Microsoft Investor Relations — fiscal Q2 2026 results page. https://www.microsoft.com/en-us/investor/earnings/fy-2026-q2/press-release-webcast

21. “$37 billion annual run rate, up 123% year-over-year” — CIO Dive coverage of Microsoft’s AI run rate growth. https://www.ciodive.com/news/microsoft-AI-cloud-growth-earnings-Q1/804338/

22. “Microsoft guided to roughly $190 billion of capital expenditure for calendar 2026” — CNBC coverage of Microsoft’s fiscal Q3 2026 earnings and 2026 CapEx outlook. https://www.cnbc.com/2026/04/29/microsoft-msft-q3-earnings-report-2026.html

23. “buys Alphabet in Q1 2023” — 24/7 Wall St. coverage of Pershing’s Q1 2023 Alphabet purchase post-ChatGPT. https://247wallst.com/investing/2026/05/15/bill-ackman-bought-google-after-the-chatgpt-panic-up-300-he-bought-amazon-after-liberation-day-up-36-now-hes-backing-up-the-truck-again/

Disclosure

This article is for educational and entertainment purposes only. Nothing here is investment advice, tax advice, legal advice, or accounting advice. I am not your CPA, accountant, attorney, or tax professional, and I am not acting as your financial advisor in this article.

Investing involves risk, including loss of principal. Past performance does not predict future results. Markets, tax law, and individual circumstances change. Before acting on anything you read here, consult a qualified professional who knows your full situation.

Position disclosure: I, members of my family, and clients of VAP Wealth Advisors may own positions in the security or securities discussed in this article. Specifically, Microsoft (MSFT) has been held in the ValueAligned Portfolio prior to and independent of Pershing Square’s February 2026 entry. I may buy, sell, add to, or trim positions at any time without prior notice and without updating this article. Nothing said here should be taken as a recommendation to buy, sell, or hold any specific security.

Views expressed are my own and do not necessarily reflect the views of VAP Wealth Advisors.

Regulatory Information

ValueAligned Partners LLC is a Registered Investment Advisor. For additional information, please refer to our Form ADV, available on the SEC’s website at www.adviserinfo.sec.gov.

David Berkowitz CIO

I’m Berk — Investor, Educator, and Owner. For 40 years I’ve helped families think like owners and invest in great companies. Earlier in my career I was head trader for a $250 million hedge fund, advised Fortune 500 boards and C-level executives and taught 10,000 of their employees at multi-billion-dollar companies, and trained non-financial employees in value-based management.

See Full Bio