OpenAI Says Anthropic Is Inflating Its Revenue by $8 Billion
Summary
A leaked internal memo from OpenAI’s chief revenue officer accuses rival Anthropic of overstating its annual run rate by roughly $8 billion through aggressive accounting, while laying out a five-priority battle plan to dominate enterprise AI. With both companies reportedly eyeing IPOs this year
OpenAI’s chief revenue officer Denise Dresser sent a four-page internal memo to employees on April 13, 2026, claiming that Anthropic’s reported $30 billion annualized run rate is inflated by approximately $8 billion due to how the company accounts for revenue-sharing arrangements with Amazon and Google, according to a report by The Verge.
The memo alleges that Anthropic grosses up cloud reseller revenue — counting the full amount before deducting the share paid to distribution partners — while OpenAI reports its Microsoft revenue share on a net basis. If the accusation holds, Anthropic’s comparable run rate would sit closer to $22 billion, placing it below OpenAI’s stated $25 billion. Neither company has commented publicly on the specific figures.
The revenue dispute is one piece of a broader competitive offensive. Dresser described the AI market as the most competitive she has ever seen and outlined five priorities for Q2: winning the enterprise model layer with OpenAI’s new “Spud” model, establishing its Frontier platform as the default agent infrastructure, expanding distribution through a recently announced Amazon partnership, selling across the full product stack, and building out a deployment engine called DeployCo.
The memo takes direct aim at Anthropic’s strategic positioning on multiple fronts. Dresser argued that Anthropic’s failure to secure sufficient compute capacity is already degrading its product experience through throttling and weaker availability. She characterized Anthropic’s developer-focused coding strength as a narrow wedge that becomes a liability as AI expands beyond engineering teams into broader enterprise workflows.
The sharpest language targeted Anthropic’s brand identity. Dresser wrote that Anthropic’s narrative is rooted in fear and restriction, contrasting it with OpenAI’s framing as a democratizing force. OpenAI CEO Sam Altman reinforced this positioning in February when he publicly stated that Anthropic serves expensive products to wealthy customers.
The accounting disagreement highlights a growing transparency problem in the AI industry. As both companies prepare for potential public offerings, investors and analysts will need to reconcile fundamentally different revenue reporting methodologies. Gross versus net recognition can dramatically alter how large a company appears, and neither approach is inherently wrong — but the gap matters when hundreds of billions in valuation hang on growth narratives.
Dresser urged employees to think of OpenAI not as a collection of separate product lines but as a platform company with multiple entry points and one integrated enterprise offering. The memo’s closing message was blunt: the market is theirs to win.