Anthropic's origin: no garage, no angels, pure venture capital

Anthropic, the ethical AI fairy tale nobody believes

Anthropic's origin is a fairy tale for adults. The official version says that a handful of OpenAI engineers, with six-figure salaries, quit their jobs due to a sudden existential concern about artificial intelligence and founded the company over Zoom, in the middle of the pandemic, with no garage involved. The legend even portrays them signing the contract on a ketchup-stained napkin. The problem: almost nobody believes it.

The protagonists were not newcomers: they earned salaries that in the real world are enough to buy a flat and keep quiet. And they weren't jumping into the void: there were venture capital funds willing to burn $20 million on the idea. The humanitarian motivation clashes with the industry's math. Not for nothing, the same speech about saving humanity was already used by Google at the turn of the century with its "Don't be evil", a motto born in 2000, popularized with the 2004 IPO and eventually diluted in the search engine's terms of use.

The Chinese alternative that breaks the board

The debate is not only about Anthropic's biography. While US labs sell subscriptions, open models of Chinese origin are distributed for free, like Deepseek. The technical discussion has turned into a practical point: with 64 gigabytes of VRAM, a mid-sized company can run local models instead of handing its data to third parties. Proprietary labs have already protested. And it's only the beginning: the war is on.

At this point, some will still believe in the enlightened founder and others will see a marketing operation with a safety net. If open models keep breaking down entry barriers, the moat of AI giants will narrow. Then it will be clear whether the ethical narrative was conviction or simply good business.