Story
August 18, 2026
Anthropic’s $65B Run Rate Puts OpenAI’s IPO Clock Under Pressure
Anthropic’s reported revenue surge has pushed it ahead of OpenAI on annualized run rate and intensified the contest to reach public markets first. Yet the headline comparison comes with a major caveat: the two AI labs may not count revenue the same way.
Anthropic’s reported revenue sprint has turned a long-running AI rivalry into a sharper financial race. The company now appears to hold the headline advantage over OpenAI—just as both groups prepare to test investors’ appetite for blockbuster IPOs.
The acceleration began from an already formidable base. Bloomberg-reported figures put Anthropic’s annualized revenue run rate at $9 billion at the end of last year, rising to $47 billion in May and above $65 billion by the end of July. Investors cited by the Financial Times expect a potential $100 billion-to-$120 billion finish for 2026.1
The quarterly picture is even more striking. Anthropic generated more than $11.5 billion in preliminary second-quarter revenue, after $4.73 billion in the first quarter—more than doubling sequentially and exceeding the year-earlier period by more than 14 times, according to documents reported by Bloomberg.2
That puts pressure on OpenAI’s own public-market narrative. OpenAI’s latest reported revenue run rate was $40 billion, up from $20 billion at the end of 2025. But neither the apparent lead nor the gap is definitive: the companies may calculate revenue differently, making a simple run-rate contest an imperfect measure of who is winning.
The next test is timing. Both companies have reportedly filed confidential IPO paperwork, while Anthropic is expected to seek a listing as soon as the fall, potentially at a valuation of $2 trillion or more. Going first could give the Claude maker a wider route to raise capital for the costly computing capacity behind its models.
Investors are also watching whether revenue quality—not just volume—becomes the deciding edge. Gavin Baker of Atreides Management said, “Anthropic was much more token efficient than OpenAI but OAI has closed some of the gap.”2 The argument is that better answers can reduce costly reruns and human checking, even where model prices are higher.
The broader market has embraced the scale of the moment. In a repost on X, Elon Musk amplified the claim that frontier-lab annual recurring revenue had surpassed the combined revenue of Windows and Office, saying the labs were “toppling the Titans.”
3