Internal memo: OpenAI Chief Revenue Officer says Anthropic is “grossing up rev share with Amazon and Google” and overstating its “run rate by roughly $8B”
OpenAI's chief revenue officer, Denise Dresser, sent a four-page memo to employees on Sunday …
The VergeHayden Field
Context & Ripple Effects
The memo arrives alongside OpenAI’s account that its Microsoft arrangement constrained access to customers using Bedrock, making cloud-distribution terms a central part of its competitive positioning rather than a back-office detail.
It also follows internal concern that Google’s AI progress could create economic headwinds for OpenAI. The new allegation therefore sharpens a contest over how to compare AI companies’ commercial scale when model providers sell through hyperscaler platforms.
First-order effects
OpenAI’s sales organization gains an internal argument for discounting Anthropic’s reported scale; Anthropic’s stated run-rate narrative faces fresh scrutiny, but the memo’s claim is an allegation rather than independently established fact.
Amazon and Google are drawn into the comparison because their revenue-share arrangements can affect whether end-customer spending is presented as provider revenue or platform-mediated revenue.
Second-order effects
Enterprise buyers and investors may place more weight on definitions—gross versus net revenue, recurring run rate, and cloud-channel economics—when comparing OpenAI and Anthropic, rather than treating headline figures as directly comparable.
Cloud partners and model providers have added incentive to clarify channel attribution and contract economics as they compete for the same enterprise workloads; OpenAI’s reported Bedrock access constraint illustrates how distribution agreements can shape that competition.
Third-order effects
If competing AI vendors continue to use different revenue-recognition presentations, market share and valuation comparisons will increasingly depend on disclosure quality and the economics retained after cloud-platform revenue shares.
The broader market may split between model makers with direct customer ownership and those whose growth is mediated by hyperscalers, making channel control as consequential as reported top-line scale.
The trend: AI competition is moving from a race over model capability toward a contest over cloud distribution, revenue attribution, and the unit economics each provider actually retains.
OpenAI (who reports/leaks revenues with their Microsoft revenue share included) accuses Anthropic of “grossing up” their runrate, inflating it by “roughly” $8bn. Cannot wait for these companies to file their S-1s. They're so mad at each other. [embedded post]
@bdsams ... Yes. SEC tends to care about comparability for direct competitors that are public but will look at the substance of the contracts. Either way, would be helpful to see apples-to-apples NET revenue to each company for comparability
At this point, I assume that every internal memo at an AI lab is just written for public release. The labs are certainly capable of keeping secrets that don't leak, so they must realize that all-hands memos or announcements just go right to The Information or Wired or whatever.
Newly leaked OpenAI memo says Anthropic's reported annual run rate of $30B is inflated by $8B. Anthropic includes gross revenue from cloud resellers (expenses their share as sales & marketing). OpenAI nets cloud sales. If correct, OpenAI at $25B run rate vs. Anthropic at $22B. [i…
Was there not a report on this a couple of weeks ago that Ant counts its revenue differently? I wouldn't trust any of these people, and especially the Effective Altruists who can justify any sin so long as they believe it results in the greater good (eventually).
OpenAI leaders should stop caricaturing Anthropic. It encourages tribalism at a time when safety cooperation is urgently needed Don't know this person so am assuming she is genuine but Anthropic's “story” is not “built on fear, restriction, and the idea that a small group of
The end result of startups counting dirty revenue is 3 “agentic commerce” companies, all founded by former HFT guys, round tripping the same $125k a hundred times per second. They will raise the first quadrillion dollar Series A
OpenAI's chief revenue officer sent a 4-page memo to employees on Sunday about the company's strategic direction, emphasizing the need to lock in users, build a moat and grow its enterprise business. (It also threw shade at its longtime rival Anthropic.) https://www.theverge.com/…
🚨OpenAI Chief Revenue Officer memo to employees LEAKED: >"Claude has become a religion, that's the level of that mania" >"Anthropic's strategy is fear, restriction, and the idea that a small group of elites should control AI" >"They made a strategic misstep to not acquire [image]
Congrats to OpenAI to sorta figuring out Anthropic's playbook one year too late. But it's still clear based on this, they don't completley understand the structural advantages. And if they think Anthropic is a “one product company”...Ooof. Hope the new podcast helps tho!
Great scoop from The Verge on The OpenAI internal memo. I scooped last week that the different accounting used by the two companies could cause up to an $8 billion discrepancy in how revenue is counted. OpenAI is sharing that number internally now. It's still true that Anthropic
is catching up. OpenAI is still chasing Anthropic on the enterprise front. But I'd argue winning consumer is the much more important battle for OpenAI. You only have one chance to win consumer AI. You will have many chances to win in the enterprise.
Holy shit OpenAI's CFO is basically saying Anthropic's stated run rate is inflated by around $8B at the current $30B ARR figure because they gross up rev share with Amazon and Google on a gross basis instead of netting out the partner cut. And that was not even the only shot. [im…