Welcome back to SavvyMonk, your one-stop for AI and tech news that actually matters. This week, a leaked four-page memo from OpenAI's chief revenue officer landed in the press, and it reads less like a strategy document and more like a company trying very hard to convince itself it's still winning.

Let's get into it.

You don’t need to be a coder to make AI work for you. Subscribe to Mindstream and get 200+ proven ideas showing how real people are using ChatGPT, Midjourney, and other tools to earn on the side.

From small wins to full-on ventures, this guide helps you turn AI skills into real results, without the overwhelm.

Get Your Free Guide

On April 13, 2026, OpenAI's Chief Revenue Officer Denise Dresser sent a four-page internal memo to staff. It was supposed to be a Q2 strategy document. But The Verge obtained it, and by Monday it was all over tech media. The reason it got so much attention is simple: a significant portion of the memo is devoted to telling OpenAI employees that their biggest rival is not as big as it looks.

Dresser came to OpenAI in December 2025 from Slack, where she was CEO. She has experience inside public company earnings cycles, and that background clearly shapes how she thinks about the revenue story both companies are building ahead of their expected IPOs.

The Core Accusation

Anthropic recently reported an annualized revenue run rate of $30 billion, up from roughly $9 billion at the end of 2025. That kind of growth, 10x year-over-year, was the number everyone in the industry was talking about. It put Anthropic ahead of OpenAI's stated run rate of around $25 billion, and that gap seems to be what prompted the memo.

Dresser's argument is that the $30 billion figure is inflated by an accounting choice. Anthropic books the full value of revenue generated through its cloud distribution partners, primarily AWS and Google Cloud, on a gross basis. That means when a customer buys Claude through AWS, Anthropic counts the entire transaction as its own revenue, including Amazon's cut.

OpenAI, by contrast, reports its revenue from the Microsoft Azure partnership on a net basis, deducting Microsoft's share before recognizing the income. "Their stated run rate is inflated," the memo reads. "They use accounting treatment that makes revenue look bigger than it is, including grossing up rev share with Amazon and Google."

OpenAI's analysis suggests this overstates Anthropic's run rate by roughly $8 billion, bringing the adjusted figure closer to $22 billion. On that basis, OpenAI is still ahead at $25 billion.

Is the Argument Valid?

Here is the important nuance. Both approaches are entirely legal under US GAAP. Neither company is doing anything wrong. The gross versus net distinction is a documented and well-understood accounting difference. The question is which treatment is more reflective of economic reality, and which one will survive the scrutiny of a public company S-1 filing.

Dresser made exactly this point: "We report Microsoft rev share net, which is more in line with standards we would be held to as a public company." Anthropic pushed back, telling CNBC that it recognizes gross revenue because it is the principal in the transaction and cloud partners serve as the distribution channel. That is a standard and accepted justification for gross recognition under accounting rules. So both sides have a defensible position. The market will ultimately decide which framing it prefers when both companies go public.

The Other Shots Dresser Took

The revenue argument was only part of the memo. Dresser also went after Anthropic on three other fronts.

On compute, she argued that Anthropic made a "strategic misstep to not acquire enough compute," and that customers are already feeling it through throttling, weaker availability, and less reliable uptime.

OpenAI, she wrote, recognized the exponential compute curve earlier and acted faster. A parallel investor memo projects OpenAI at 30 gigawatts of compute capacity by 2030, compared to 7 to 8 gigawatts for Anthropic by end-2027. Anthropic has countered with a Broadcom and Google TPU deal worth 5 gigawatts starting 2027, and disputes the framing.

On product breadth, Dresser acknowledged that Anthropic's early focus on coding tools gave it an early wedge with enterprise customers. Ramp data across more than 50,000 US businesses shows Anthropic at 30.6% of enterprise AI-paying customers, closing in on OpenAI's 35.2%. But Dresser's argument is that a coding-first identity becomes a liability as AI expands beyond developer teams. "You do not want to be a single-product company in a platform war," the memo states.

On brand positioning, she took a sharper tone: "Their story is built on fear, restriction, and the idea that a small group of elites should control AI." This echoes language used by US AI Czar David Sacks, and it is notable that OpenAI is borrowing that political framing inside an internal memo.

The Spud Model and Five Q2 Priorities

Dresser also outlined OpenAI's strategic priorities for the quarter. The most interesting one is a new model codenamed "Spud." She described it as an important step in the intelligence foundation for the next generation of work, with early customer feedback pointing to stronger reasoning, better understanding of user intent, and more reliable production performance. She did not give a release date, and the name is almost certainly a codename rather than a product name.

The five Q2 priorities she outlined were: winning the enterprise model layer with Spud, establishing the Frontier platform as the default infrastructure for AI agents, expanding distribution through a recently announced Amazon partnership, selling across OpenAI's full product stack, and building out a deployment engine called DeployCo.

The Amazon partnership detail is worth noting. Earlier in the memo, Dresser blamed OpenAI's tie to Microsoft as something that limited our ability to meet enterprises where they are. The new Amazon deal is clearly meant to address that constraint.

The Bottom Line

This memo got leaked, and that is almost certainly not an accident. It reads like a document meant to travel. The accounting argument against Anthropic is real and will matter when both companies file for public offerings, likely later this year.

But the memo also reveals that OpenAI is watching Anthropic's trajectory closely enough to send a four-page counter-narrative to its own employees. That is not the posture of a company that feels comfortable in its lead.

Category: Competitive Intelligence

"You are a senior business analyst. I am going to give you two companies in the same industry. Compare their revenue reporting methodologies, identify any accounting differences that could affect headline figures, and explain which approach is more conservative and why. Company A reports revenue [describe A's method]. Company B reports revenue [describe B's method]. Summarize the key differences in plain language and flag which treatment public market investors are more likely to prefer."

The interesting thing about this memo is not the accounting argument. It is that OpenAI felt it needed to make one. When you are writing a four-page internal document to explain why your competitor's number is not real, you are telling your employees something you probably did not intend to say. The race is closer than the scoreboard suggests, and everyone inside OpenAI knows it. Hit reply, I read every response.

See you in the next one.

— Vivek

P.S. If you know someone who follows the AI business closely or works in enterprise tech, forward this their way. They can subscribe at https://savvymonk.beehiiv.com/