Microsoft's Anthropic Bet Pays Off Big in One Quarter, While OpenAI Investment Stumbles

In its fiscal Q4 2026 earnings, Microsoft revealed a striking divergence in the performance of its two major AI lab investments. The company recorded a $3.2 billion gain on its stake in Anthropic, boosting diluted earnings per share by 33 cents. This comes less than a year after Microsoft invested $5 billion in Anthropic as part of a complex agreement that also saw the AI lab commit to purchasing $30 billion worth of Azure services.

Meanwhile, Microsoft's investment in OpenAI—where it owns about 27%—took a $600 million write-down in the same quarter, reducing EPS by about 7 cents. However, on a full-year basis, the OpenAI investment generated a $5 billion gain, contributing $0.67 to EPS. The contrast is striking: one quarter of Anthropic gains nearly matched an entire year of OpenAI gains, a fact Microsoft chose to disclose prominently.

For developers and startups watching the AI infrastructure economy, this news underscores how intertwined the fortunes of AI labs and cloud providers have become. Microsoft's deal-making with both Anthropic and OpenAI reflects a deliberate strategy to hedge its bets across competing foundational model vendors, while ensuring Azure remains the default compute platform for both. The huge Azure consumption commitments from these labs are a double-edged sword: they lock in revenue, but they also expose Microsoft to the financial health of its partners.

While the $600 million write-down is a rounding error for a company that posted $90 billion in quarterly revenue, the signal matters. OpenAI's valuation and revenue growth may be facing headwinds, or Microsoft is simply marking the investment conservatively. Either way, this quarter's numbers offer a rare glimpse into the real economics of AI's biggest power players.