Microsoft’s latest earnings revealed a striking split in its AI bets: a $3.2 billion quarterly gain from its Anthropic investment, versus a roughly $600 million markdown on OpenAI. In one quarter, Anthropic delivered nearly as much as OpenAI did all year. It’s a rare public glimpse into how Microsoft’s rival AI stakes are actually performing.

Key Takeaways

• Microsoft logged a $3.2 billion quarterly gain from Anthropic
• It marked down its OpenAI investment by about $600 million
• Anthropic’s quarterly gain nearly matched OpenAI’s full-year gain
• Microsoft invested $5 billion in Anthropic in November 2025
• The disclosure signals a shift in how Microsoft weighs the two labs

What Microsoft Reported

The detail was tucked into strong results. In its fiscal fourth-quarter earnings for the year ending June 30, Microsoft recorded its Anthropic investment as a $3.2 billion gain, boosting diluted earnings per share by 33 cents, against reported diluted EPS of $4.81 for the quarter.

The Anthropic stake is recent. Microsoft invested $5 billion in Anthropic in November 2025 as part of a circular agreement under which the AI lab also agreed to buy $30 billion worth of Azure services.

OpenAI told a different story. Microsoft said its OpenAI investment did not fare nearly as well and marked it down about $600 million, reducing diluted EPS by roughly 7 cents, on the roughly 27% stake it owns in the ChatGPT maker.

The Striking Comparison

The one-quarter-versus-full-year gap stands out. For the fiscal year, Microsoft’s OpenAI investment generated a $5 billion gain and added $0.67 to EPS, meaning Anthropic delivered nearly as much in a single quarter as OpenAI did across the whole year.

Microsoft found it noteworthy enough to flag. The company doesn’t routinely update the value of its Anthropic stake each quarter, so choosing to disclose the $3.2 billion gain was itself a signal of how significant the figure was.

The accounting differs between the two. Microsoft discusses its OpenAI investment every quarter but accounts for the value of the stake rather than reporting the revenue-share payments it receives, so these figures reflect investment value, not cash in hand.

Context Behind the Numbers

The OpenAI markdown was minor for Microsoft. Despite the $600 million reduction, the write-down was mostly a rounding error for a company that reported $90 billion in revenue and $35.8 billion in net income for the quarter.

The broader results were strong. Microsoft’s cloud business crossed $100 billion in annual Azure revenue, and the company cited both the Anthropic gain and lower-than-expected costs from its first voluntary retirement program as tailwinds, though its Xbox unit took an impairment charge.

These are paper gains, not cash. The figures reflect changes in the accounting value of Microsoft’s equity stakes, not dividends or direct returns, an important distinction given the circular nature of many AI investment deals.

Why It Matters

The split reveals a portfolio strategy paying off. Rather than betting everything on OpenAI, Microsoft has spread its AI stakes, and this quarter Anthropic’s rising value cushioned a decline in OpenAI’s, exactly the diversification benefit a portfolio is meant to provide.

It hints at shifting fortunes. Anthropic outperforming OpenAI on Microsoft’s books, even for one quarter, suggests the AI race is less settled than headlines imply, notable coming from the company funding both sides.

The disclosure choice speaks volumes. By breaking from its usual practice to highlight the Anthropic gain, Microsoft signaled that its relationship with the two labs is evolving, a theme that runs through its increasingly competitive posture toward both. For investors, it’s a rare window into which AI bets are actually delivering.

Digital Trendings is your trusted source for AI news and updates, stay tuned for more.