“Google is a core AI winner: hyperscaler scale, best-in-class TPUs, and a dual Nvidia/TPU strategy create a durable cost and customer advantage.”
Why he says it — point by point
FAVORABLE & AGAINST · BOTH KEPTWhy
I think that some of the other hyperscaler candidates, Google in particular, SpaceX in particular, are better positioned for the future and are better bets on growing well and making money in as hyper in the hyperscaler side of the business.
Google's been doing it longer. Tesla's been doing it a long time with its inference chips and Dojo.
Google probably more than anyone else has access to just a ridiculous number of customers.
they are probably the biggest hyperscaler. They're really good at it.
Google has been making TPUs. They're on the eighth generation TPU inference and and training versions.
Google's TPUs are considered very very good. Google is also one of the largest customers of Nvidia. They they play both ends. They know how to play both bets. Some of their customers want to use Nvidia hardware. Some of their customers want to use their TPUs. Their TPUs give them a cost advantage.
it may be something that's helping them remove the infection and that's able to help Google operate more effectively.
Risk
Google has the bureaucracy problem. Google has the like the woke infection.
The structured call
The receipt
Publish-day price $$356.13 · the claim is anchored to the moment it was said.