Six European offices in under a year, a $965B valuation, and now the keys to Elon's supercomputer—Anthropic is done asking permission to compete with Google and OpenAI.
The Summary
- Anthropic raised $65B at a $965B valuation, putting it within striking distance of becoming the first AI company to hit $1T without being owned by a tech incumbent
- Milan marks the sixth European city where Anthropic has opened offices in under a year, signaling enterprise demand outside the U.S. is real and accelerating
- Anthropic secured a six-month lease of SpaceX's Colossus 1 compute cluster, giving Claude temporary access to one of the world's largest GPU arrays while it builds its own infrastructure
- The compute lease, the capital raise, and the geographic sprint point to the same conclusion: Anthropic is moving from scrappy challenger to global infrastructure player
The Signal
Anthropic's $65B raise at a $965B valuation is the kind of number that rewrites what "AI startup" means. For context, that's more than the GDP of most countries. It's also more cash than Meta raised in its entire history before going public. The capital is earmarked for compute expansion, research, and enterprise adoption. Translation: Anthropic is betting it can build the infrastructure to train and serve models at Google and Microsoft scale without being Google or Microsoft.
The Milan office is the tell. Six European cities in under a year isn't a PR tour. It's distribution. Enterprise customers in regulated industries—banking, healthcare, manufacturing—want AI that doesn't route through U.S. hyperscalers. Anthropic is giving them a local alternative with Claude, and those customers are writing checks big enough to justify boots on the ground in Dublin, Paris, London, and now Milan.
"Enterprise demand outside the U.S. is no longer a future story. It's current revenue."
Here's where it gets interesting. Anthropic leased SpaceX's Colossus 1 cluster for six months. Colossus 1 is one of the largest GPU arrays on the planet, purpose-built for xAI's Grok. Anthropic doesn't need it forever. They need it now, while their own infrastructure catches up to the capital they just raised. This is the AI equivalent of renting a stadium while you build your own. It's also a signal that the compute bottleneck is real enough that competitors are willing to share hardware.
The lease deal also tells you something about xAI's positioning. If Elon is willing to rent out Colossus for six months, it means xAI either has excess capacity or is confident enough in its own roadmap that sharing compute with a rival doesn't scare them. Either way, it's a data point on how fast infrastructure is being built and how fluid the competitive landscape still is.
Key dynamics at play:
- Anthropic is racing to build global enterprise presence before OpenAI or Google lock in long-term contracts
- The $965B valuation assumes Claude becomes the default model for regulated industries that can't or won't use hyperscaler AI
- Six-month compute leases from competitors suggest the infrastructure market is still fragmented enough that short-term deals make sense
The Implication
If you're building agents or automation tools, watch where Anthropic plants offices next. Those cities are where enterprise budgets for AI are already flowing. If you're an enterprise buyer, the fact that Anthropic is willing to open local offices means they're serious about compliance, data residency, and long-term support in a way that API-first competitors aren't.
The trillion-dollar valuation threshold matters because it changes Anthropic's leverage. At $965B, they can negotiate compute deals, poach talent, and undercut hyperscaler pricing in ways that weren't possible two years ago. For builders in the agent economy, this means one more credible infrastructure alternative to OpenAI and Google. For enterprises, it means pricing pressure and more negotiating power. Watch who wins the next wave of Fortune 500 AI contracts. That will tell you if Anthropic's sprint was worth it.