Anthropic at $350 billion: when caution becomes the most expensive commodity in the world

Remember when Anthropic was known as "the startup that's afraid of AI"? A group of OpenAI alumni left Sam Altman to build safe models free from corporate profit-seeking. Forget that. Today, Anthropic is discussing a $350 billion valuation. That is a lot of money by any measure, and it puts the company on par with the market capitalization of the world's largest corporations, which took decades to build. If the deal goes through, Dario Amodei's company will officially cement its status as the leading contender for the throne in the large language model world.

The mechanics are simple: Anthropic is preparing a tender offer. It lets employees sell their shares and turn "paper" wealth into real money. Companies usually do this to retain top talent — when an engineer sees seven zeros in their banking app, they're far more willing to stay at the office until midnight. But the scale is striking. Alongside the employee share sale, the company plans to raise about $20 billion in fresh capital. For comparison, that exceeds the annual research budgets of many developed countries.

Why now? The answer is Claude. The third version of the model didn't just catch up with GPT-4 — in some areas, such as coding and handling huge context windows, it pulled ahead. The enterprise sector, long wary of OpenAI's unpredictability, saw Anthropic as "the adult in the room." Amazon and Google have already poured billions into the company, and now they appear ready to double down. For the cloud giants, Anthropic functions less as a startup and more as critical infrastructure that guarantees their servers will stay loaded with compute for years to come.

This deal can't be separated from what's happening at the competition. While OpenAI goes through internal conflicts and a change in its organizational structure, Anthropic shows remarkable stability. The company is methodically building what it calls "Constitutional AI," and investors are willing to pay a huge premium for that predictability. $350 billion is the price of trust that AI can be controlled. Or, more likely, it's the price of the fear of missing out on carving up the biggest pie in tech history.

But a valuation like this also brings enormous responsibility. At $350 billion, Anthropic can no longer afford to be a "lab." It is now a profit machine. Every Claude update will be judged not by the quality of its poetry but by how many billions it saved or earned for Fortune 500 clients. Dario Amodei has long avoided publicity and the image of a "tech prophet," but with this much money behind him, he'll have to step into the spotlight. The only question is whether there will still be room for the very "safety" that started it all when sums of this magnitude are at stake.

The bottom line: the era of romanticism in AI is officially over. Anthropic is now a full-fledged financial leviathan. Will it live up to investor expectations, or are we watching the biggest bubble since the dot-com era?