Anthropic метит в $350 миллиардов: Claude стоит дороже вашего воображения

It seems that in the world of large language models, we've entered an era where a billion dollars has stopped being a serious figure. Anthropic, a company founded by former OpenAI employees, is now discussing a valuation of 350 billion dollars (3500). If you thought OpenAI's 157 billion was a lot, then Dario Amodei and his team decided to prove that investor appetites have no limits.

This isn't just about injecting fresh capital, but about a large-scale restructuring of ownership where employees will be able to cash out their shares against the backdrop of record interest in the Claude model. The context of this deal looks even more ironic if you remember the company's founding story. Anthropic was created as a "safe" alternative to OpenAI, focusing on ethics and so-called constitutional AI.

However, the reality is that training models at the level of Claude 3.5 Opus requires colossal computational power and even more colossal amounts of money. To compete with Microsoft and Google, it's not enough to be just "good guys"—you need access to a checkbook that doesn't run out.

The current 20 billion dollar round (200) is an attempt to cement the status of main competitor to GPT-4 while other players try to catch up with the departing train. Why are investors willing to pay such money right now? The answer lies in performance.

Recent versions of Claude have shown that Anthropic is capable of producing a product that works more stably and logically than OpenAI's solutions. Among developers, Claude 3.5 Sonnet has become the de facto standard for coding and complex text analytics.

While Sam Altman is busy with media projects and global alliances, Amodei methodically improves the architecture, proving that a focus on quality can be worth more than aggressive marketing. The market believes that Anthropic is not just a "number two," but a potential leader of the next technological cycle. It's interesting that part of the deal involves secondary sales of shares by employees.

This is an important signal for the entire industry. When a startup allows its people to exit in cash at such a valuation, it means the company is confident in its future and not afraid of talent drain. It's also a way to retain the best talents who could leave for their own projects.

On Silicon Valley's talent market, there's currently a real war underway, and the opportunity to become a millionaire "on paper," turning it into real money, is the best argument in the competition for world-class engineers. However, one should ask how justified such multiples are. A valuation of 350 billion dollars puts Anthropic in the same league as the world's largest corporations, which have real revenue, factories, and millions of paying customers.

AI startups so far only have promises to change the world and electricity bills from Nvidia. If this bubble bursts, it will drag the entire technology sector with it. But while money flows like rivers, no one wants to be the one who turns off the lights in the server room last.

The belief that AGI will justify any investment remains the main driver of modern economics. The bottom line: Anthropic is definitively transforming from a scientific project into a financial monster. Will the company be able to maintain its principles of "safe AI" under pressure from investors expecting a 350 billion dollar return?