Microsoft disclosed a $3.2B pre-tax gain on its Anthropic investment during Q4 of fiscal year 2026 (the quarter ending June 30), adding 33 cents to diluted earnings per share, according to TechCrunch and the company’s earnings presentation.
In the same quarter, Microsoft marked down its OpenAI investment by approximately $600M, reducing diluted EPS by about 7 cents.
The Asymmetry
The divergence between the two AI investments is notable. Microsoft’s full-year OpenAI gains totaled $5B, adding 67 cents per share across all four quarters, according to the company’s SEC filing. A single quarter of Anthropic returns ($3.2B) came close to matching that entire annual figure.
Microsoft invested $5B in Anthropic in November 2025 as part of a circular arrangement: Anthropic committed to purchasing $30B in Azure services. Microsoft owns approximately 27% of OpenAI and receives revenue-share payments, though it does not disclose the amounts.
Microsoft does not routinely update the value of its Anthropic stake each quarter. It does report OpenAI investment values quarterly. The fact that Microsoft chose to disclose the Anthropic gain this quarter, as TechCrunch noted, “is so noteworthy that Microsoft disclosed it.”
The Hedge Structure
The financial results illuminate how Microsoft has positioned itself in the AI lab competition. The company holds significant stakes in both Anthropic and OpenAI. Both labs are contractually obligated to purchase Azure infrastructure. Whether Claude or GPT wins enterprise market share, Microsoft collects compute revenue from both.
The Q4 FY2026 results also underscore how AI lab valuations are becoming material contributors to Microsoft’s earnings. On total Q4 revenue of $90B and net income of $35.8B, the $3.2B Anthropic gain and the $600M OpenAI write-down represent meaningful line items, not rounding errors. Microsoft reported $331.8B in revenue and $133.7B in net income for the full fiscal year, with diluted EPS of $17.95.
The Incentive Question
For enterprises choosing between AI platforms, the financial arrangement raises a structural question: Microsoft is simultaneously a cloud infrastructure provider, an investor in competing AI labs, and an increasingly direct competitor building its own models and agent tools. It profits from both Anthropic and OpenAI succeeding, from both consuming Azure compute, and from building its own alternatives.
That triple position is unusual in technology. It is more common in financial services, where investment banks routinely advise competing parties in the same market. Whether enterprises treat Microsoft as a neutral infrastructure provider or a strategically conflicted participant may depend on how transparently the company manages these overlapping interests.