Artificial intelligence is not cheap, and Anthropic knows this well. The company founded by former OpenAI members needs a $200 billion injection to keep competing. Google, its main backer, has already put money on the table, but the total figure reveals just how much the race to dominate AI requires resources on the scale of a small country.
The cost of training a model that doesn't hallucinate 💸
Developing models like Claude requires GPU clusters, giant data centers, and enough electricity to light up half a city. Anthropic has prioritized safety and interpretability, which further drives up the cost. Google isn't funding this out of philanthropy; it's seeking a solid counterweight to OpenAI and Microsoft. The investment covers everything from custom chips to researcher salaries, in a war where hardware defines who leads.
Google digs into its pockets and hopes not to find them empty 😅
At this rate, Google will have to sell its own cloud to keep Anthropic afloat. The $200 billion figure sounds like a bad joke, but it's real. Meanwhile, Anthropic's engineers ask for more GPUs, and OpenAI's engineers laugh with their Microsoft deals. In the end, AI will be so smart that it will know exactly how much money has been spent on it. And it will probably have a good laugh.