google.com, pub-8701563775261122, DIRECT, f08c47fec0942fa0
USA

Nvidia Shares Drop on Report of Google Challenge in AI Chips

Nvidia Corp. its shares fell on a report that Meta Platforms Inc. was in talks to spend billions of dollars on Google’s artificial intelligence chips; This shows the internet search leader is making progress in its efforts to rival the industry’s best-selling AI accelerator.

Meta is in talks about using Google chips, known as tensor processing units, or TPUs, in data centers in 2027, The Information reported, citing an unidentified person familiar with the talks. The news outlet also noted that Meta may lease chips from Google’s cloud division next year.

Most Read from Bloomberg

The deal will help create TPUs as alternatives to Nvidia chips, which are the gold standard for big tech firms and startups from Meta to OpenAI that need computing power to develop and run AI platforms.

Nvidia’s shares fell as much as 2.7% in after-hours trading. Google owner Alphabet Inc. gained 2.7%, adding to recent optimism about the latest version of its Gemini AI model.

Google previously signed a deal to supply up to 1 million chips to Anthropic PBC, highlighting the potential for long-term challenges to Nvidia’s dominant market position.

Following the announcement of the Anthropic agreement, Seaport analyst Jay Goldberg called it “really strong validation” for TPUs. “A lot of people were already thinking about it, and a lot more people are probably thinking about it now,” he said.

Meta representatives declined to comment, while Google did not immediately respond to requests.

What Does Bloomberg Intelligence Say?

Meta’s possible use of Google TPUs currently used by Anthropic suggests that third-party vendors offering large language models are likely to use Google as a secondary accelerator chip supplier for inference in the near term. By our calculation, Meta’s capex of at least $100 billion for 2026 suggests it will spend at least $40-50 billion on inference chip capacity next year. Consumption and accumulation growth on Google Cloud may accelerate compared to other hyperscalers and neo-cloud peers due to demand from enterprise customers looking to deploy TPUs and Gemini LLMs on Google Cloud.

– Mandeep Singh and Robert Biggar, analysts

Click here for research.

Alphabet-related Asian stocks rose in early trading in Asia on Tuesday. IsuPetasys Co., which supplies multilayer sheets to Alphabet in South Korea. It reached a new intraday record, up 18%. In Taiwan, shares of MediaTek Inc. rose almost 5%.

A deal with Meta, one of the world’s biggest spenders on data centers and artificial intelligence development, would represent a win for Google. But much depends on whether tensor chips can demonstrate the power efficiency and computing power needed to become a viable option in the long term.

First developed more than 10 years ago specifically for AI tasks, the tensor chip is gaining momentum outside its own company as a way to train and run complex AI models. Its appeal as an alternative has grown at a time when companies around the world are worried about over-reliance on Nvidia in a market where even Advanced Micro Devices Inc. is a distant second.

Graphics processing units, or GPUs, part of the chip market dominated by Nvidia, were created to speed up the rendering of graphics (mostly in video games and other visual effects applications), but they turned out to be well-suited for training artificial intelligence models because they can handle large amounts of data and calculations. TPUs, on the other hand, are a type of specialized products known as application-specific integrated circuits or microchips designed for a separate purpose.

Tensor chips have also been adapted as accelerators for AI and machine learning tasks in Google’s own applications. Because Google and its DeepMind unit have developed cutting-edge AI models like Gemini, the company has been able to learn from these teams, right down to chip designers. At the same time, the ability to personalize chips has also benefited AI teams.

–With help from Riley Griffin and Carmen Arroyo.

Most Read from Bloomberg Businessweek

©2025 Bloomberg LP

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button