Microsoft records $3.2B Anthropic gain while OpenAI stake loses $600M
On July 29, 2026, Microsoft disclosed sharply different results from its two largest AI investments for fiscal Q4, which ended June 30. Anthropic generated a $3.2 billion accounting gain and added $0.33 to diluted earnings per share, while OpenAI produced a $600 million write-down that reduced EPS by about $0.07.
Why the divergence matters
These movements are not the same as operating revenue or cash received from customers. They show how changes in the assessed value and performance of strategic holdings can influence reported profit, even when the companies involved compete in the same market.
For Microsoft, the OpenAI decline was small relative to quarterly revenue of $90 billion and net income of $35.8 billion. Yet the contrast exposes the volatility embedded in large AI partnerships and matters to investors, cloud customers and software providers building around either lab.
The investments behind the figures
Microsoft invested $5 billion in Anthropic in November 2025 through an arrangement that also committed Anthropic to purchasing $30 billion of Azure services. Microsoft does not routinely disclose an updated value for this holding every quarter, making the $3.2 billion gain an unusual data point.
Microsoft owns about 27% of OpenAI and also receives undisclosed revenue-share payments. Although the stake lost $600 million in value during the quarter, it generated a $5 billion gain over the full fiscal year and added $0.67 to annual EPS. Microsoft reported full-year revenue of $331.8 billion, net income of $133.7 billion and diluted EPS of $17.95.
The contrast adds a financial dimension to Microsoft's sales positioning across OpenAI and Anthropic, because Microsoft is simultaneously an investor, infrastructure supplier and commercial partner to competing AI labs.
For businesses, the practical lesson is to separate model quality, provider economics and infrastructure dependence when selecting an AI platform. A diversified supplier strategy can reduce exposure to valuation swings, changing commercial terms and shifts in the competitive balance between labs.