
Google turns up the heat on Nvidia as AI chip competition intensifies
Google is expanding its custom AI chip lineup, a move that could ratchet up pressure on Nvidia while the broader AI trade shows signs of strain. The search giant plans to launch two new tensor processing units (TPUs) next year: one built for confidential computing to serve customers handling sensitive data, and another designed for high-volume workloads such as coding and drug discovery.
Bloomberg Intelligence projects the TPU hardware-and-software market could grow from roughly $12 billion this year to more than $135 billion by 2028. Its analysts also estimate Google’s chips will cost about 30% less than Nvidia’s GPUs—an advantage that could squeeze Nvidia’s pricing power and chip market share.
The timing is awkward for Nvidia. The stock plunged after the company disclosed its board has opened an investigation into reports that CEO Jensen Huang had been targeted by foreign intelligence, wiping out more than $250 billion in market value at one point on Monday. Nvidia shares are still off about 4.3% even as the Nasdaq rallied over 2% and Google gained roughly 3.4%.
The turbulence comes amid a broader cooling in AI-related names: neocloud and infrastructure stocks have tumbled, with firms such as IREN (IREN $21.78), Cipher Digital (CIFR $6.8150) and CoreWeave (CRWV) trading as much as 70% or more below their highs.
Why crypto watchers should care
The GPU market matters to crypto because many miners, cloud-rendering services and Web3 compute projects rely on Nvidia’s GPUs for heavy workloads. Lower-cost TPUs from Google could shift enterprise demand for AI compute, potentially easing GPU supply strains or pressuring prices—factors that ripple into mining economics and the broader tokenized-compute ecosystem.
Bottom line: Google’s TPU expansion adds a new competitive front to the AI chip race, and combined with the unexpected governance probe at Nvidia, it creates a volatile backdrop for hardware suppliers, cloud players and crypto-adjacent businesses dependent on high-performance GPUs.
Google is expanding its custom AI chip lineup, a move that could ratchet up pressure on Nvidia while the broader AI trade shows signs of strain. The search giant plans to launch two new tensor processing units (TPUs) next year: one built for confidential computing to serve customers handling sensitive data, and another designed for high-volume workloads such as coding and drug discovery.
Bloomberg Intelligence projects the TPU hardware-and-software market could grow from roughly $12 billion this year to more than $135 billion by 2028. Its analysts also estimate Google’s chips will cost about 30% less than Nvidia’s GPUs—an advantage that could squeeze Nvidia’s pricing power and chip market share.
The timing is awkward for Nvidia. The stock plunged after the company disclosed its board has opened an investigation into reports that CEO Jensen Huang had been targeted by foreign intelligence, wiping out more than $250 billion in market value at one point on Monday. Nvidia shares are still off about 4.3% even as the Nasdaq rallied over 2% and Google gained roughly 3.4%.
The turbulence comes amid a broader cooling in AI-related names: neocloud and infrastructure stocks have tumbled, with firms such as IREN (IREN $21.78), Cipher Digital (CIFR $6.8150) and CoreWeave (CRWV) trading as much as 70% or more below their highs.
Why crypto watchers should care
The GPU market matters to crypto because many miners, cloud-rendering services and Web3 compute projects rely on Nvidia’s GPUs for heavy workloads. Lower-cost TPUs from Google could shift enterprise demand for AI compute, potentially easing GPU supply strains or pressuring prices—factors that ripple into mining economics and the broader tokenized-compute ecosystem.
Bottom line: Google’s TPU expansion adds a new competitive front to the AI chip race, and combined with the unexpected governance probe at Nvidia, it creates a volatile backdrop for hardware suppliers, cloud players and crypto-adjacent businesses dependent on high-performance GPUs.




