Anthropic Passed OpenAI on Valuation. “Worth” Doesn’t Mean What It Used To.

Anthropic is now technically worth more than OpenAI. That sentence has two words doing enormous work: “technically” and “worth.” Bloomberg and TechCrunch both report that Anthropic is weighing a $50 billion raise at a valuation between $850 billion and $900 billion — which would clear OpenAI’s last post-money figure of $852 billion. Three months ago, in February 2026, Anthropic was worth $380 billion. The number more than doubled in a quarter, and the thing that changed was not the product.

What actually changed is how Anthropic is being used — as an instrument, not just a company. The valuation game at this scale isn’t a report card on current revenue. It’s leverage. A near-trillion-dollar number tells enterprise customers you’re not going anywhere. It tells the cloud hyperscalers bankrolling you that your independence is worth protecting. It sets the IPO floor. TechCrunch’s sources describe investors being given a “48-hour allocation deadline,” which is either a sign of genuine scarcity or manufactured urgency. Probably both. Anthropic’s safety-first positioning has always been an enterprise differentiator; a near-trillion valuation converts that positioning into a balance sheet argument. The raise is as much a marketing event as a financial one.

There’s a structural story underneath the headline number that deserves more attention. Amazon has committed up to $25 billion to Anthropic; Google up to $40 billion — a combined $65 billion from the two largest cloud incumbents on the planet. Anthropic’s much-publicized independence from OpenAI’s orbit, its argument that it’s building AI differently, rests almost entirely on two companies that would be directly harmed if Anthropic disrupted the cloud market in any meaningful way. The independence is real in the sense that neither company owns Anthropic outright. It is not real in the sense that Anthropic cannot operate at this scale without them. That’s a subtle but consequential distinction when the entire brand is built on doing this differently.

This round — if it closes — would reportedly be Anthropic’s final private fundraise before a potential IPO as early as October 2026. That timeline clarifies what the raise is actually for. It isn’t R&D. It isn’t compute expansion. It’s balance sheet inflation: making the company look worth what the IPO roadshow will claim before institutional investors get to weigh in on whether that’s true. OpenAI’s IPO is still coming, and its target is around $1 trillion. Anthropic just moved the goalpost first. The scoreboard advantage is real and also temporary.

The counterargument that genuinely holds up doesn’t start with earnings — it starts with what Anthropic could become. If the company achieves anything close to safety-aligned AGI, $900 billion looks cheap. If artificial general intelligence becomes the operating layer of the global economy, you price it the way you price water rights: not by what it costs to build, but by what happens if you don’t control it. That’s the logic underlying every number at this end of the AI market, and it’s also the logic that could justify any number at all — which should give everyone involved some pause.

The Anthropic story since this site started tracking it has been one continuous ratchet: each valuation milestone surfaces the same question and nobody answers it. What does it mean to be worth nearly a trillion dollars when you haven’t gone public, when your two largest funders are also your primary distribution channels, and when your brand rests entirely on safety being a reason not to rush? The mission and the valuation are awkward roommates, and they’re about to share a bunk with an IPO roadshow. That should be the most interesting part of what happens next.

Sources: TechCrunch · Bloomberg · CNBC