Even after absolutely destroying expectations with its latest earnings report (and proving, once again, that the “AI bubble” crowd might need a new hobby), competitors and anxious investors are still grasping at any headline they can find to try and knock Nvidia off its throne. 👑

And once again, they got one.

Nvidia (NVDA) shares slipped in premarket trading on Tuesday after reports that Meta may start using Google’s custom AI chips (specifically tensor processing units or TPUs) in its data centers beginning in 2027. Meta is also reportedly considering renting Google’s TPUs from their cloud unit starting next year. Basically, Meta seems to be shopping around… and the markets reacted immediately. 📉

Google (GOOGL), on the other hand, is loving the attention. Alphabet shares spiked more than 6% on Monday and were trading another 3% higher Tuesday morning. And Broadcom, which helps design Google’s TPUs, also enjoyed an extra 2% bump after surging 11% the day before. Not bad for one announcement.

Google’s TPUs aren’t new, but they’ve gotten very good. Introduced in 2018 for internal use, they’re now powerful, specialized chips built specifically for AI tasks. And because TPUs are custom–built, they give Google a notable advantage. If Meta actually adopts them, it would be a major validation for Google’s chip strategy and a meaningful hit to Nvidia’s absolute dominance. 

Nvidia, however, isn’t exactly nervous. After all, their GPUs remain the backbone of the global AI buildout, and their lead is nowhere near collapsing. Still, Big Tech has been desperately trying to diversify supply chains, mostly because everyone is tired of waiting in line for Nvidia’s sold-out chips. Which is… fair. 🤷‍♂️

Many argue that the endless debate over whether we’re in an “AI bubble” and whether tech valuations are stretched. And it just so happens that Nvidia has been at the heart of that whole conversation, when investors are emotional and apparently eager to panic over any sign of competition.