The memo doesn’t read like corporate blandishment. It reads like a war plan drafted by someone who knows the enemy is at the gates.

“The market is as competitive as I have ever seen it,” OpenAI chief revenue officer Denise Dresser wrote in a four-page internal memo sent to staff on Sunday, viewed by The Verge. The language throughout is blunt, urgent, and mercifully free of the usual Silicon Valley equivocation. The core message: OpenAI must build walls around its business before competitors — chiefly Anthropic — catch up.

Dresser, who recently absorbed duties from former COO Brad Lightcap as he transitions to a special projects role, framed the challenge in explicitly platform terms. “Multi-product adoption makes us harder to replace,” she wrote. “We should stop thinking like a company with separate product lines. We should think like a platform company with multiple entry points and one integrated enterprise offering.”

That single sentence is the strategy. Everything else is execution.

The Anthropic Offensive

The memo’s most striking passage is a point-by-point takedown of Anthropic, OpenAI’s closest rival in the enterprise AI market. Dresser accused Anthropic of inflating its stated run rate by roughly $8 billion through accounting treatment that grosses up revenue shared with Amazon and Google. “We report Microsoft revshare net, which is more inline with standards we would be held to as a public company,” she wrote — a line that reads differently given that both companies reportedly plan to go public this year.

She also attacked Anthropic’s compute strategy as a “strategic misstep” resulting in throttling and weaker availability for customers, and dismissed its coding-focused product as too narrow. “You do not want to be a single-product company in a platform war.”

Anthropic’s narrative, according to Dresser, is “built on fear, restriction, and the idea that a small group of elites should control AI.” OpenAI’s counter-message, she wrote, is to “build powerful systems, put in the right safeguards, expand access, and help people do more.”

Both companies are racing toward the same destination: the public markets, where these competing narratives will be tested by investors rather than journalists.

Outgrowing Microsoft

The memo’s candor extends to OpenAI’s relationship with Microsoft, its largest investor and computing partner. Dresser described the partnership as “foundational” but conceded it has “limited our ability to meet enterprises where they are — for many that’s Bedrock,” Amazon’s cloud platform for AI models.

Since announcing the Amazon partnership in late February, Dresser wrote, inbound demand has been “frankly staggering.” The Amazon Stateful Runtime Environment — which enables memory and continuity across interactions — gives OpenAI a way to reach AWS-native customers, strengthen its position with regulated buyers, and integrate more deeply into production workflows.

According to MarketWatch, Dresser’s memo called the Amazon alliance the key to unlocking AI monetization in the enterprise market. That framing signals how far OpenAI is willing to stretch beyond Redmond to chase revenue.

Stargate UK Shelved

The strategic retrenchment extends to infrastructure. OpenAI confirmed last Thursday it is pausing Stargate UK, a multibillion-pound data centre project at Cobalt Park in North Tyneside, north-east England. The project, part of a wider £31 billion UK tech investment package announced with considerable fanfare, would have deployed thousands of Nvidia chips through a partnership with Nscale.

An OpenAI spokesperson said the company would proceed only when “the right conditions such as regulation and the cost of energy enable long-term infrastructure investment.” The BBC reported that concerns include uncertainty over whether the UK will change copyright law to allow AI training on copyrighted works — a proposal that drew loud opposition from artists including Elton John.

The UK government struck a diplomatic note, saying it had attracted more than £100 billion in private AI investment and would “continue to create the right conditions.” But the optics are unwelcome for a government that has championed AI as an economic growth engine.

Follow the Money

Read together, the memo and the paused deal tell a coherent story. OpenAI is pulling back on capital-intensive infrastructure bets in uncertain markets while concentrating fire on the thing that matters most: converting its early lead in AI into defensible, recurring enterprise revenue before the field catches up.

Dresser’s five priorities — win the model layer, win the agent platform, expand through Amazon, sell the full stack, and own deployment — are not aspirations. They are instructions. The flywheel she describes is simple: better models drive usage, usage drives integration, integration drives multi-product adoption, and multi-product adoption raises switching costs.

That is how you build a moat. The question is whether the walls go up fast enough.

Sources