Google is ready to drop $40B on Anthropic. Here’s what that means.

Google is reportedly preparing to sink up to $40 billion into Anthropic — a mix of cash and compute credits — as the AI arms race shifts from model quality to raw infrastructure. That’s a staggering number, even by Big Tech standards.

For context, Microsoft has already poured something like $13 billion into OpenAI, and Amazon has committed $4 billion to Anthropic. Google’s potential $40B bet makes those look like pocket change. The question is: why?

The short answer is compute. Training and running frontier models eats up GPUs like nothing else. Google has its own TPUs, but even they can’t keep up with demand. By throwing compute credits at Anthropic, Google isn’t just buying equity — it’s locking up capacity for itself. This is a play for scarcity, not just returns.

Then there’s the timing. This news comes right after Anthropic quietly released its Mythos model, which is heavily focused on cybersecurity. Mythos isn’t a general-purpose chatbot like Claude. It’s designed for threat detection, vulnerability analysis, and automated response. The limited release suggests Anthropic is being cautious, which is smart given how dangerous a powerful security model could be if mishandled.

I’ve seen a few takes calling Mythos “the Stuxnet of LLMs.” That’s dramatic, but not entirely wrong. A model that can autonomously find and exploit zero-days is a dual-use nightmare. Google’s investment might be a way to steer Anthropic’s priorities toward defensive applications rather than offensive ones. Or maybe they just want first dibs on the tech.

Either way, the $40B figure is eye-popping. If it goes through, it would be one of the largest private investments in AI history. But I’m skeptical about the compute portion. Google has been tightening access to its TPU clusters, and handing over that much capacity to an external partner could strain internal projects. There’s a reason OpenAI relies on Azure — Microsoft has the cloud infrastructure to spare. Google might be betting that Anthropic’s models will be worth the resource drain.

What’s more interesting is the strategic positioning. Google already has DeepMind, which is arguably the strongest pure research lab in the world. Why double down on Anthropic? Because DeepMind is focused on fundamental breakthroughs, while Anthropic is shipping products. Google needs both to compete with OpenAI’s ChatGPT and Microsoft’s Copilot ecosystem.

So this isn’t just about money. It’s about securing compute, talent, and market position in a landscape that’s shifting faster than most people realize. The Mythos release is a signal that Anthropic is moving beyond safety research into applied AI. And Google wants a front-row seat.

If I had to guess, we’ll see more of these mega-deals in the next 12 months. Compute is the new oil, and the companies that control it will shape the next decade of AI. $40B sounds insane today, but in five years, it might look like a bargain.