Google’s AI Chips Rocked AMD and Nvidia. Analysts Say “We Don’t Think Meta’s Aspirations Are Changing.”

Alphabet’s (GOOGL) push into selling its own AI hardware set off one of the most dramatic pivots in the semiconductor sector this week. A report that Google is in talks to sell its tensor processing units to Meta Platforms (META) sparked immediate pressure on Nvidia (NVDA) and Advanced Micro Devices (AMD), two companies that have defined the modern AI boom. Traders, however, appear to be recalibrating their initial fears.

Shares of Nvidia fell 2.6% on Tuesday and AMD dropped 4.2% as investors reacted to the idea that Google’s specialized chips could chip away at the massive GPU demand that has driven both stocks to historic levels. By Wednesday morning, the shock had cooled. Nvidia reversed about half its slide, rising 1.3%, and AMD gained 2.7%. For now, the market appears to be weighing the scale of Google’s ambitions against the size of the AI wave that both GPU makers continue to ride.

Google Tests Rival Chip Ambitions

Google’s potential plan involves supplying Meta with tensor processing units starting in 2027. These TPUs are application-specific integrated circuits built for highly optimized AI workloads and can offer a more energy-efficient and cost-effective option compared with general-purpose GPUs. The Information reported that Meta is exploring these chips as part of a broader effort to diversify its data-center infrastructure.

The strategic logic is straightforward. Meta needs cheaper compute at massive scale, and Google is one of the few companies with the engineering depth to offer an alternative. But even if TPUs gain traction, analysts argue that the GPU market is far too large, and too essential to current AI workflows, for Google to meaningfully undercut Nvidia or AMD in the near term.

Analysts Reassess Nvidia and AMD

Morgan Stanley (MS) analysts Brian Nowak and Joseph Moore directly addressed the market’s anxiety. “We don’t think Meta’s aspirations are changing: they want to diversify their compute infrastructure,” they wrote. This line reframed Tuesday’s selloff as an overreaction to a development that aligns with long-running trends in cloud architecture.

The analysts added that return on investment could ultimately determine how compute spending evolves, but they do not expect a sudden transition. Nvidia, in particular, remains tied deeply into Google’s own AI buildout. The company supplies Google with around twenty billion dollars worth of hardware, reinforcing a relationship that boosts both sides even as Google develops its own ASICs.

Nvidia echoed this sentiment. “We’re delighted by Google’s success – they’ve made great advances in AI and we continue to supply to Google,” a company spokesperson told Barron’s. “Nvidia offers greater performance, versatility, and fungibility than ASICs, which are designed for specific AI frameworks or functions.”

AMD Faces the Tougher Narrative

AMD is positioned more precariously in the eyes of investors. The company has spent the past two years crafting a story around becoming the primary alternative to Nvidia, especially as a major supplier to OpenAI. Monday’s report complicated this narrative as Google’s Gemini models continue to compete with OpenAI’s technology, creating questions about where AMD fits in a multi-platform AI world.

Bernstein analyst Stacy A. Rasgon summarized the concern. “The company has built their story around becoming the viable second source to NVDA, and hitched their wagon to OpenAI’s horse,” he wrote. The introduction of Google TPUs into Meta’s future plans adds a new axis of competition that investors will have to map carefully. Bernstein (AB) kept its Outperform rating on Nvidia and a Market Perform on AMD.

Investors can compare AI stocks side-by-side based on various financial metrics and analyst ratings on the TipRanks Stocks Comparison Tool. Click on the image below to find out more.