Nvidia Stock Dips as Meta Reportedly Opts for Google AI Chips

Shares of Nvidia (NVDA) experienced a downturn Tuesday, triggered by a report from The Information suggesting that Meta Platforms (META) is evaluating the potential adoption of Google’s internally developed Tensor Processing Units (TPUs).

Nvidia’s stock dipped 4% in premarket trading, reflecting investor concern over the potential loss of a major client. Conversely, Alphabet (GOOGL), Google’s parent company, saw its shares rise 4.2%, building on a 6% surge from the previous day. This divergence highlights the market’s perception of Google’s gain at Nvidia’s expense.

The report indicates that Meta is considering deploying Google’s TPUs in its data centers as early as 2027. Furthermore, Meta is reportedly exploring the option of renting TPU capacity from Google Cloud starting next year, signifying a potentially significant shift in its AI infrastructure strategy.

“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,” stated a Google spokesperson, underscoring the company’s dual-pronged approach to the burgeoning AI infrastructure market.

Google’s foray into AI-specific hardware began in 2018 with the introduction of its first-generation TPU, initially designed for internal use within its cloud computing operations. Since then, Google has consistently refined and advanced its TPU architecture to address the demanding requirements of contemporary artificial intelligence workloads.

TPUs, as application-specific integrated circuits (ASICs), offer a number of advantages and disadvantages to the general-purpose GPUs offered by NVIDIA. Their custom silicon design enables superior performance and energy efficiency for specific AI tasks, such as inference. However, that specialization also means that TPUs may lack the flexibility of GPUs when it comes to supporting new and evolving AI models.

Meta’s potential utilization of TPUs would represent a significant validation of Google’s long-term investment and strategic vision in vertically integrated AI solutions.

Broadcom (AVGO), a key partner in the design and manufacturing of Google’s TPUs, saw its shares increase by over 2% in premarket trading, following an 11% jump on Monday. This upward momentum underscores the symbiotic relationship between Google’s AI ambitions and its hardware partners.

Nvidia currently dominates the AI semiconductor market with its high-performance Graphics Processing Units (GPUs). These GPUs have become essential components within the accelerating AI infrastructure development across industries. While Nvidia’s market leadership remains secure in the short-term, the increasing presence of Google’s TPUs contributes to a more competitive landscape in the burgeoning AI silicon market.

A key driver behind the diversification trend is the desire among AI infrastructure builders to reduce their reliance on a single dominant supplier like Nvidia, mitigating supply chain risks and fostering greater negotiating leverage.

Meta stands out as a major investor in AI infrastructure, projecting capital expenditures between $70 billion and $72 billion for the current year, demonstrating its commitment to advancing its AI capabilities across its diverse portfolio of products and services. This is an expensive undertaking for the company, and they are looking for anyway to control costs.

These market fluctuations occur amid ongoing discussions concerning a potential “AI bubble” and highly valued tech companies. The industry remains cautious regarding the sustainability of hypergrowth.

Nvidia has been at the center of this debate. Despite reporting an optimistic sales outlook for the current quarter last week, the tech stock market experienced a decline following the announcement. This suggests that uncertainty exists concerning market valuations and the future growth trajectory of companies that are perceived to significantly profit from AI.