Microsoft Records a $3.2 Billion Anthropic Gain and a Quarterly OpenAI Write-Down

Microsoft recorded a $3.2 billion quarterly gain on its investment in Anthropic, adding approximately 33 cents to diluted earnings per share. During the same period, the company reduced the carrying value of its OpenAI investment by roughly $600 million, subtracting about seven cents per share.

The Anthropic position comes from a $5 billion investment announced in November 2025 as part of a broader agreement that included a commitment by the AI company to buy $30 billion of Azure services. Microsoft does not update that valuation on a fixed quarterly schedule, making the disclosed gain particularly notable.

Microsoft owns about 27% of OpenAI and also receives revenue-sharing payments, although it does not disclose the amount. The quarterly reduction looks less concerning over a full year: Microsoft reported that the OpenAI investment produced a $5 billion annual gain and added 67 cents to earnings per share.

Both positions are small relative to Microsoft's overall results. The company generated $90 billion in quarterly revenue and $35.8 billion in net income. Their strategic value may be greater than the accounting impact because the relationships drive demand for Azure and give Microsoft access to leading models.

The reported gains also depend partly on private-market marks rather than cash realized through a sale. An IPO or acquisition could validate them, while a weaker transaction could reverse part of the accounting benefit. Microsoft has the balance sheet to tolerate that volatility, but smaller investors should not treat an internal valuation change as equivalent to operating profit.

The accounting contrast illustrates why strategic investments can make quarterly earnings harder to interpret. A gain on Anthropic may reflect a higher private valuation, while the OpenAI relationship includes operating commitments and write-downs that move differently. Investors should separate these changes from Azure demand, software subscriptions and cash flow. Microsoft can justify investments if they secure cloud workloads and product access, even when a single quarter looks uneven. The important question is whether the portfolio strengthens the platform at an acceptable cost, not which private-company mark produces the most attractive headline in one report.

The comparison should not be read as a simple judgment that one laboratory is performing better. Private-company valuations, investment terms and accounting treatment can move differently from operating results. It does show that Microsoft has turned model competition into a financial portfolio, gaining influence and cloud commitments even when the companies compete with one another and with Microsoft itself.