OpenAI Internal Memo Attacks Anthropic Revenue Claims in Corporate AI War

OpenAI's chief revenue officer fired the first public shot in what is becoming an unusually open corporate war. In a four-page internal memo sent to staff on Sunday, Denise Dresser laid out a detailed case against Anthropic's competitive positioning — questioning its revenue figures, its compute infrastructure, and its public messaging strategy.

The $8 Billion Revenue Accounting Dispute

The centerpiece of Dresser's memo is an accounting argument. Anthropic publicly cites a $30 billion annual revenue run rate, which has been widely reported as evidence of the company's rapid growth. Dresser argues that figure is overstated by roughly $8 billion because of how Anthropic counts revenue from its cloud distribution partners.

Anthropic books revenue from AWS, Microsoft Azure, and Google Cloud on a gross basis — recording the full amount billed through those channels, including the cloud providers' cut. OpenAI reports its Microsoft Azure revenue on a net basis, deducting the partner's share. Both approaches are permissible under US GAAP, but the choice materially affects headline numbers. Under Dresser's analysis, Anthropic's real run rate is closer to $22 billion — slightly below OpenAI's reported $25 billion figure.

Distancing from Microsoft, Embracing Amazon

The memo also marks a notable shift in OpenAI's public posture toward Microsoft. Dresser wrote that the Microsoft deal "limited our ability" to reach clients using AWS Bedrock — a significant admission given how central the Microsoft partnership has been to OpenAI's enterprise strategy. The memo explicitly touts the company's expanded Amazon partnership as the vehicle to recapture that ground.

Messaging and Compute Criticisms

Beyond the financials, Dresser took aim at Anthropic's brand positioning: "Their story is built on fear, restriction, and the idea that a small group of elites should control AI." She also pointed to what she framed as a structural disadvantage — claiming Anthropic failed to secure enough compute early, which she says shows up as throttling, weaker availability, and unreliable uptime for customers.

"You do not want to be a single-product company in a platform war," the memo reportedly concluded — a dig at Anthropic's narrower product portfolio compared to OpenAI's expanding ChatGPT ecosystem.

The Bottom Line

The memo leaked almost immediately, which is either a sign of poor internal culture or a deliberate signal. Either way, it reveals how intensely OpenAI is tracking Anthropic's growth narrative and how worried it is about losing enterprise deals to its fastest-growing rival. The revenue accounting dispute may be technically valid — but the fact that OpenAI felt the need to make it in a memo to its own staff suggests the competitive pressure is real.