Microsoft's Anthropic Investment Outperforms OpenAI in Q4 2026

The Numbers: Anthropic vs OpenAI, Side by Side

Here's the comparison that made analysts sit up:

• •Anthropic: a $3.2 billion gain, adding 33 cents to diluted earnings per share
• •OpenAI: marked down by roughly $600 million, cutting diluted EPS by about 7 cents

To put the Anthropic number in perspective, Microsoft's CFO Amy Hood pointed out that the gain exceeded the operating income of Microsoft's entire More Personal Computing segment — Windows, devices, Xbox, and search advertising — by around 16%. That's a paper gain from a minority stake outperforming an entire consumer product division.

Zooming out to the full fiscal year tells a more balanced story, though: Microsoft's OpenAI investment generated a $5 billion gain over the full fiscal year, adding 67 cents to annual EPS, and Microsoft holds roughly a 27% stake in OpenAI. So while Anthropic won the quarter, OpenAI still delivered more total value across the year — the Q4 story is about momentum and direction, not total size.

Why the Anthropic Stake Moved So Much

A key detail explains why this number stood out at all: Microsoft doesn't routinely revalue its Anthropic investment every quarter, unlike its OpenAI stake, which it updates and discusses each quarter. When a company chooses to disclose an infrequent revaluation, it's usually because the number is too significant to bury.

The investment itself dates back to late 2025: Microsoft invested $5 billion in Anthropic in November 2025, as part of a deal in which Anthropic also agreed to purchase $30 billion worth of Azure services. That structure — an investment paired with a large committed cloud spend — is often called a "circular" deal in the industry, since the invested capital effectively flows back to the investor through service purchases.

This wasn't an isolated Microsoft move either. Nvidia also committed up to $10 billion to Anthropic around the same time, with Anthropic services becoming available on Microsoft Azure and Claude models accessible through Microsoft Foundry. That deal made Anthropic backed by all three major hyperscalers, following earlier investments from Google (roughly $3 billion for about a 14% stake) and Amazon (up to $8 billion).

The OpenAI Side of the Story

It's worth being fair to the OpenAI side of the ledger too. The $600 million write-down doesn't mean the OpenAI bet is failing — it means one quarter's mark-to-market valuation moved against Microsoft, likely reflecting normal fluctuations in how a private company's stake gets valued between reporting periods.

And the long-term OpenAI relationship remains enormous in scale: Microsoft has invested more than $13 billion in OpenAI in total, giving it a 27% stake, and it retains 80% of sales revenue when it sells OpenAI models to Azure customers — a materially better revenue-sharing arrangement than what it holds with Anthropic. Microsoft's Anthropic-related spending was reported at approaching $500 million a year, and the company has started letting Azure sales staff count Anthropic model sales toward their quotas the same way OpenAI sales already count.

So while Anthropic delivered the bigger quarterly headline number, OpenAI remains the deeper, more commercially integrated partnership for Microsoft.

Why This Matters Beyond One Earnings Call

1. Diversification is paying off

Microsoft betting on two competing AI labs simultaneously — rather than going all-in on one — meant that when one investment had a rough quarter, the other cushioned the blow at the balance-sheet level.

2. Anthropic's commercial momentum is real

Anthropic's Claude assistant and Claude Code have grown more popular this year, increasing the company's computing needs — demand strong enough that Anthropic committed to paying $1.25 billion per month for computing power through May 2029. Growing usage is generally what drives the kind of valuation jump reflected in Microsoft's gain.

3. The AI investment landscape is getting more competitive, not less

With Google, Amazon, Nvidia, and Microsoft all holding Anthropic stakes, and Microsoft also deeply invested in OpenAI, the era of a single "AI partner" per hyperscaler is over. Companies are hedging across labs.

4. Circular deals are becoming the norm

The investment-for-cloud-spend structure behind the Microsoft-Anthropic deal is becoming a template across the industry, and it's worth watching how regulators and investors respond to that pattern going forward.

What to Watch Next

• •Whether Microsoft begins revaluing its Anthropic stake more frequently, given how material this one update was
• •How the Microsoft-Anthropic AI chip discussions (reportedly around Microsoft's Maia processors) progress
• •Whether Anthropic's growing compute costs — including its multi-year deal for computing power — affect future valuation swings
• •Whether OpenAI's next quarterly mark reverses this quarter's write-down

Conclusion

Microsoft's fiscal Q4 2026 earnings revealed something more interesting than just strong quarterly numbers — a real-time comparison of how its two competing AI bets are performing. Anthropic's $3.2 billion gain outpaced OpenAI's quarterly write-down, driven by strong product momentum and a cloud-spend-linked investment structure. But zoomed out to the full year, OpenAI still delivered more total value and remains the deeper commercial partnership. The bigger takeaway: Microsoft's strategy of backing multiple AI labs, rather than betting on just one, is starting to show its logic on the balance sheet.