Nvidia fell more than four percent in morning trading after reports emerged that Mark Zuckerberg’s Meta is in talks to spend billions on Google’s AI chips, potentially establishing them as a viable alternative to Nvidia’s industry-leading offerings.
Bloomberg reports that Nvidia shares are taking a hit following reports that Mark Zuckerberg’s Meta is engaged in discussions to invest billions of dollars in Google’s AI chips. Known as tensor processing units (TPUs), these chips are poised to challenge Nvidia’s long-standing dominance in the AI accelerator market.
According to a report from the Information, which cited an unnamed source familiar with the matter, Meta is considering using Google’s TPUs in its data centers starting in 2027. The social media giant is also reportedly exploring the possibility of renting chips from Google’s cloud division as early as next year. This news caused Nvidia’s shares to slide by more than four percent in morning, while Google shared nudged slightly positive.
If the deal materializes, it would mark a significant milestone for Google’s TPUs, establishing them as a credible alternative to Nvidia’s chips. Currently, Nvidia’s offerings are considered the gold standard for major tech companies and startups, including Meta and OpenAI, which rely on their computing power to develop and run advanced artificial intelligence platforms. Google has already made inroads in this space, securing a deal to supply up to 1 million TPUs to Anthropic. However, Nvidia still maintains a dominant position in the market.
Seaport analyst Jay Goldberg praised the potential Meta deal as a “really powerful validation” for Google’s TPUs, suggesting that it could prompt more companies to consider them as a viable option. Representatives from Meta declined to comment on the matter, while Google did not immediately respond to requests for comment.
Despite the potential win for Google, the long-term success of its TPUs will depend on their ability to demonstrate superior power efficiency and computing performance compared to Nvidia’s offerings. The tensor chip, which Google first developed over a decade ago specifically for AI tasks, has been gaining traction outside the company as a means to train and run complex AI models. This growing interest comes at a time when companies worldwide are increasingly concerned about overreliance on Nvidia, with even AMD which is trailing far behind in the market.
While graphics processing units (GPUs), Nvidia’s area of expertise, were initially designed to accelerate graphics rendering for applications like video games, they have proven to be well-suited for training AI models due to their ability to handle vast amounts of data and computations. In contrast, TPUs are a type of application-specific integrated circuit (ASIC) designed for a specific purpose.
Breitbart News recently reported that Nvidia posted strong sales of its GPUs and related hardware as Jensen Huang attempts to assuage market fears of an AI bubble:
For the quarter, Nvidia reported adjusted earnings per share of $1.30, surpassing the estimated $1.25. Revenue came in at $57.01 billion, beating the $54.92 billion consensus estimate. The company also provided a robust sales guidance for the current quarter, projecting about $65 billion in sales, compared to the $61.66 billion expected by analysts.
Nvidia’s success is largely attributed to the insatiable demand for its AI chips, known as GPUs, which are used by leading tech companies such as Microsoft, Amazon, Google, Oracle, and Meta to develop new AI models. The company’s sales and outlook serve as a crucial indicator of the health of the AI boom in the technology industry.
Nvidia CEO Jensen Huang addressed concerns about an AI bubble, stating, “From our vantage point, we see something very different.” The company’s most important business segment, data center sales, reported $51.2 billion in revenue, easily surpassing analyst expectations of $49.09 billion and representing a 66 percent year-over-year increase. Within data center sales, $43 billion was attributed to the company’s GPUs, with much of the growth driven by initial sales of the GB300 chips. Networking, which allows multiple GPUs to work as a single computer, accounted for $8.2 billion in data center sales.
Read more at Bloomberg here.
Lucas Nolan is a reporter for Breitbart News covering issues of free speech and online censorship.
Read the full article here