Google and Anthropic now officially announcing that new cloud deal worth tens of billions of dollars for access to custom AI accelerator chips.

McKenzie Segalo says more on what it means for the cloud landscape in today's tech check.

Morning, Mac.

Hey, good morning, Carl.

This is Anthropic betting big on Google, committing to use up to a million of Alphabet's custom-built AI chips and a deal that

by industry standards could be worth around $50 billion.

Now, this build adds more than a gigawatt of compute by 2026.

That's enough to power 750,000 homes.

And yes, OpenAI has announced future commitments worth 33 times that, but this deal, it is locked

in, coming online next year, and tied to explosive enterprise demand.

Claude now powers more than 300,000 businesses with revenue approaching $7 billion this

year.

Now what's also at stake here is the unusually complicated dynamic between Anthropic,

Google, and Amazon.

So by deepening its ties to Google, Mizuho says that Anthropic is signaling concern that

Amazon's in-house silicone hasn't kept pace, calling it a competitive win for Google,

and a sign that the balance of power in AI infrastructure may be shifting.

And it comes as that cloud race heats up. Amazon is still the leader here, but its dominance

is no longer what it once was. Microsoft's enterprise momentum is accelerating thanks to OpenAI,

and now Google and even Oracle are gaining ground off the back of that OpenAI Halo 2.

Still, Amazon has poured $8 billion into Anthropic, built Claude's custom supercomputer,

and it remains its primary cloud provider.

...