Anthropic Surges Past OpenAI: The New AI Leader’s Wall Street Moment

Anthropic is planning to go public as early as September. In our last briefing, we talked about its secret AI researcher automating breakthroughs; today, we see the result. The company has overtaken OpenAI in annualized revenue for the first time, pulling in 74.3 billion dollars to OpenAI’s 41.3 billion. That’s a 33-billion-dollar gap. The student has officially surpassed the master. Here's the thing about that number. It isn't just a scoreboard victory. It’s the sound of an entire industry re-aligning in real time. For years, the story was OpenAI and then everyone else. Not anymore. Data from just this month shows Anthropic gaining 1.1 percentage points of market share among business users. OpenAI? They gained just 0.23 points. That’s not a trend.

That's an ARRIVAL. And while Anthropic was surging, OpenAI was stumbling. The company has been bleeding leadership. Fidji Simo, the former Instacart CEO brought in as Sam Altman’s second-in-command, is out. Chief Revenue Officer Denise Dresser, gone after just eight months. Former COO Brad Lightcap, also departed. You don't lose that much senior talent when things are going well. Even CEO Sam Altman admitted it, telling TIME, "We clearly had some missteps as a company." He cited problems with product direction and pretraining research. That’s the CEO admitting the core engine and the roadmap were both off course, just as his main rival was preparing to ring the bell on Wall Street. So what was the misstep? They missed the real battleground.

It stopped being about who had the smartest model and started being about who owned the workflow. Specifically, the developer's workflow. A company called Cursor built a popular code editor that used AI models to help write software. It became indispensable. Instead of trying to beat Cursor, Anthropic and OpenAI are now trying to BECOME Cursor. Both companies just released their own direct competitors: Anthropic with Claude Code, and OpenAI with a new version of Codex. And according to TIME, they are luring users with heavily discounted bundles of their best models. This is a classic platform war. It's a price war designed to achieve one thing: lock-in. They don't just want you to call their API anymore. They want to own the entire screen where you work.

They want to be the operating system for creation, and they are willing to burn billions in cheap model access to buy that real estate. This isn't a product launch. It's a land grab. And it's a direct threat to every company building tools on top of their models. The platform is now competing with its own ecosystem. But you should NEVER count OpenAI out. While the headlines focused on departing executives, OpenAI was preparing its answer. It’s a new family of AI models called Astra. And from the early demos, it’s a fundamental leap. This isn't just a bigger, better GPT. This is an entirely different kind of AI. In one demonstration shown to major customers, they gave Astra a research-level math problem. The system didn't just solve it.

It split into SIXTEEN different AI agents, each tackling a piece of the problem, and then they assembled a proposed proof together. In another demo, Astra edited documents and spreadsheets across different desktop applications at full speed. It was acting like a human assistant, not a chatbot. Sam Altman predicts Astra will be "the first model where the model actually invents new things in a way that matters." That’s the bet. While Anthropic is winning the business war today with practical tools and a clear path to an IPO, OpenAI is betting it can change the rules of the game entirely with a true invention engine. They’re responding to a business problem with a science project. A massive, unbelievably ambitious science project.

That September IPO is the starting gun for the next phase of this race. Anthropic will be the first pure-play AI lab to test the public markets, trading private capital for public scrutiny. This isn't a friendly rivalry between research labs anymore. It's a fight for market share, for quarterly earnings, and for the right to define the next decade of technology. The days of moving fast and breaking things are over. The days of moving fast and reporting to shareholders are about to begin.