Nvidia stock drops as report suggests Meta could shift toward Google's AI chips

Would validate Google as a competitor in the AI hardware race

Nvidia shares slid this week after a report indicated Meta Platforms is weighing a significant shift in its long-term chip strategy, taking on Google’s chips in its datacentres instead of Nvidia’s.

Nvidia’s stock fell as much as 7% in early trading, before trimming losses to close down roughly 4%.

Meanwhile, Alphabet - Google's parent company - rose more than 4%, building on a strong rally a day earlier.

The market reaction followed a report from The Information claiming Meta is considering deploying Google's Tensor Processing Units (TPUs) in its datacentres starting in 2027, and may even rent TPUs from Google Cloud as early as next year.

"Google Cloud is experiencing accelerating demand for both our custom TPUs and NVIDIA GPUs; we are committed to supporting both, as we have for years," a Google spokesperson told The Information.

Google TPUs

Google introduced its first-generation TPU in 2018 for internal use but has since built advanced versions tailored for AI workloads.

These custom chips, designed with efficiency in mind, have become a selling point for Google Cloud customers seeking alternatives to Nvidia's GPUs.

For years, TPUs were available only through Google Cloud's infrastructure. Recently, however, Google has begun pitching TPUs for deployment inside customers' own datacentres, targeting companies such as Meta and major financial institutions.

If Meta adopts TPUs at scale, it would represent one of Google's biggest validation wins in the AI hardware race.

Meta is among the world's biggest AI infrastructure spenders, projecting $70-$72 billion in capital expenditures this year.

A partial shift toward Google's TPUs could meaningfully diversify its hardware base and reduce dependence on Nvidia's tightly constrained GPU supply.

Nvidia's competitive pressure

Nvidia's dominance in AI server chips has made it the world's most valuable company, turning massive GPU demand into record-shattering revenue and giving CEO Jensen Huang ample cash to invest in leading AI labs.

Huang has been aggressively counter-marketing against TPUs, holding deep-dive technical reviews of Google's progress and pre-emptively offering strategic deals to companies considering TPU adoption.

After Google agreed to provide up to 1 million TPUs to AI firm Anthropic, Nvidia responded by committing billions of dollars in new investments in the AI startup, securing a pledge from Anthropic to use Nvidia GPUs.

And when reports emerged that OpenAI planned to rent TPUs from Google Cloud, Huang struck a tentative deal to invest up to $100 billion so OpenAI could build its own datacentre infrastructure, including leasing Nvidia GPUs.

"We are always looking for ways to partner more deeply with leading AI labs," an Nvidia spokespeople said.

Nvidia last week reported a stronger-than-expected sales forecast, although tech shares broadly declined on valuation worries.

Even so, few analysts expect Nvidia's leadership to be meaningfully challenged in the near term.

Computing says:

The industry urgently needs an Nvidia competitor. The chip-maker is the only firm profiting from the so-called AI boom, and evidence is coming to light that it may be artificially propping up the market using some creative – though legal – practices.

A real competitor would address Nvidia’s current monopoly, lower prices and diversify the market risk, cushioning a potential bubble burst scenario.