Anthropic raises $65B in Series H funding at $965B post-money valuation
That's a lot of AGI, with some substantial firepower:
Anthropic has raised $65 billion in Series H funding led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, valuing the company at $965 billion post-money.
From Coatue, who lead the round:
...what’s changed since three months ago (the last round was only three months ago, believe it or not)?
Three things:
We’ve seen the shift from co-pilots that help make us more efficient to agents actually doing long-form work tasks on our behalf. This has meant an explosion in both productivity and ROI for customers.
Anthropic introduced Mythos Preview, a whole new model class that demonstrates a step change in intelligence level as seen in the METR chart below. It not only shows dramatic improvement to coding and long-running agents, it is also defining a new frontier for security.
And finally —> use cases spreading from coding to “everything else.”
Replies
I've tried them all for various coding tasks and most have been a frustrating waste of time. Anthropic works in a way that not only produces working code but also uses "best practice" forward looking development techniques that minimizes error detection & rewrite efforts but also suggests and tracks road map features writes user guides from the code and follows that up with tag lines, marketing content etc. Great product to work with.
$65B is a crazy number, but the bigger story is customer ROI.
The moment AI starts completing work instead of assisting with work, budgets get a lot easier to justify. That's probably why we're seeing capital flow into companies pushing agent capabilities the hardest.
Curious to see how much of this growth comes from coding versus the "everything else" category over the next 12 months.
From my opinion:
The co-pilot to agent shift is the real story here, the valuation is just a reflection of that. When AI goes from "helps me write" to "actually completes the workflow," the ROI calculus for enterprise changes completely, and so does the willingness to pay. The part I'm watching is the "use cases spreading to everything else" point from the Coatue note. Coding was the easy proof of concept because the feedback loop is fast and the output is measurable.
I believe the hard part is regulated industries like finance, healthcare, compliance, where the same agent behavior needs audit trails and governance baked in, not added later. That's where the real scale is, and also where most deployments still fail before they reach production.
Almost a trillion dollar valuation for a company that is still figuring out its business model. Meanwhile solo founders are out here building profitable products with zero funding and getting ignored. I am not anti-Anthropic — I literally use Claude every day to build — but the disconnect between how much money flows into AI infrastructure versus how much reaches the people actually making useful things with it is absurd. The tools are democratizing but the capital is not.
نتابع وننتظر الاحدث