Nvidia publicly reacted to a report suggesting Meta might switch part of its AI infrastructure to Google's TPUs, causing a drop in Nvidia's stock. The move highlights a growing rivalry, as Google’s chips gain recognition as a viable alternative to Nvidia's GPUs.

Nvidia asserts its GPUs are a generation ahead of Google’s AI chips, even as concerns arise from a potential Meta-Google deal involving tensor processing units. The company emphasizes its chips' flexibility and performance advantages compared to Google’s application-specific chips. Nvidia maintains over 90% market share in AI chips, despite increasing competition from Google’s TPUs.

Global AI computing capacity is increasing rapidly, doubling every seven months. NVIDIA dominates the market with over 60% of total compute, while Google and Amazon follow. The data is based on sales figures and financial reports, with significant growth noted since 2022.

The article discusses how companies like Anthropic are moving away from reliance on Nvidia for AI chips, exploring partnerships with Amazon and Google to broaden their hardware options. This shift is driven by tighter compute availability and the need to hedge against risks associated with a single vendor. As alternatives improve, a multi-chip market is emerging.

Google is shifting its strategy by offering its custom TPUs for deployment in customer data centers, moving away from using them only in its own cloud. Meta is reportedly in talks to integrate these chips, planning a multibillion-dollar investment starting in 2027 while also renting TPU capacity from Google Cloud. This could significantly boost Google's presence in the AI chip market and challenge Nvidia's dominance.

OpenAI has announced that it currently has no plans to deploy Google's tensor processing units (TPUs) for its products, despite early testing. This statement comes after reports indicated that OpenAI was considering using Google's AI chips to meet increasing demand.