Anthropic Just Became the Most Valuable AI Company. The Story Isn't What You Think.

Most people still think AI means ChatGPT. They use it to write emails, summarize documents, plan trips. That's the AI everyone knows.

Last week, the AI company they don't know became the most valuable in the world.

Anthropic raised a $65 billion Series H at a $965 billion valuation. That's roughly $113 billion more than OpenAI's last valuation. The maker of Claude is now sitting at the top of the AI pyramid — quietly, with most consumers still asking "Anthropic who?"

Here's why that matters.

The race that's been going on publicly — ChatGPT vs Gemini vs Copilot, "which AI writes the best email" — that's the entry-level game. Important, sure. But not where the real money is being placed.

The race that decided this valuation is different. It's about who powers actual work. Coding. Agents. Multi-step reasoning. The stuff that doesn't make for fun viral demos but quietly reshapes how things get built.

Anthropic's revenue run rate is now $47 billion. Most of that is driven by Claude Code — an AI assistant for developers that has, for many of us, stopped feeling like a tool and started feeling like a coworker who never sleeps.

I've used Claude Code every day for the last six months. It's the first AI tool I've used that I miss when I don't have it. Not because it writes flashy answers, but because it shows up reliably, reads context properly, and finishes what it starts. That's a different kind of AI than the chatbot you use to draft a poem.

And it's that kind of AI that just got valued at nearly a trillion dollars.

What this changes

Two things, mainly.

First, the public conversation about AI is mostly tracking the wrong thing. The "AI war" you read about in newspapers is about consumer chatbots. The actual war is about whether AI can do real work — code, reasoning, agents, decisions. Anthropic just made the case that the work-AI category is where the money is.

Second, OpenAI now has a real competitor that isn't just keeping up. For most of the past three years, the narrative was "OpenAI leads, everyone else follows." That narrative is broken. The follower is now the leader, at least by valuation.

What I think it means

The next decade of AI won't be decided by who writes the best email. It'll be decided by who can be trusted to do real work — work that has consequences if it goes wrong.

If you've been judging AI by what your kids show you on TikTok, you've been watching the wrong screen.

Open Claude. Try Claude Code. Watch the gap between "AI that's fun" and "AI that's useful." That's the gap that just got priced at a trillion dollars.