Business
Meta in Talks to Use Google’s Chips, Nvidia Shares Drop 3%
New York, NY – Nvidia‘s shares fell by 3% following reports that Meta is in negotiations to move part of its infrastructure to Google‘s custom-made chips, a potential deal worth billions. Analysts express concerns this shift could significantly disrupt the tech landscape.
Brent D. Griffiths wrote that it has been a ‘rocky couple of weeks’ for Nvidia, as uncertainties around the company and the AI market continue to arise. Investors are reacting to these developments with some trepidation, indicating that ‘lingering doubts’ about Nvidia’s position in the market are creeping back in.
Dina Bass noted that customers have been calling for ‘more competition’ in the chip market, suggesting that alternatives may be ‘hiding in plain sight’. Google’s tensor processing units (TPUs) were engineered a decade ago and are recognized for their efficiency in generating AI responses compared to Nvidia’s graphics processing units (GPUs).
Despite being ‘less adaptable’ than GPUs, Google’s chips are cheaper and more energy-efficient. Reports indicate that major tech firms are now keen to adopt TPUs to meet the increasing demands for computing power.
Google’s recent announcements have ‘pierced Nvidia’s aura of invulnerability’, making it a fierce competitor to the once-dominant chipmaker. Analyst reports highlight that Google’s chips could be up to ten times less expensive than Nvidia’s equivalents, posing a notable challenge for Nvidia.
While Nvidia insisted it remains ‘a generation ahead’ of competitors and the only platform supporting all AI models effectively, the looming threat from Google is significant. Eva Roytburg commented that Nvidia’s actions to defend its market position could indicate that it feels ‘spooked’ by the developments.
The proposed agreement between Meta and Google is not finalized, leaving room for changes. Ongoing discussions may not guarantee a finalized deal, according to The Wall Street Journal. Both companies are actively trying to attract clients through various financing options to entice the rollout of their chips.
As the AI landscape evolves, questions remain regarding how Meta plans to utilize these chips, either for training AI models or enhancing query response efficiency.
Bloomberg states that, despite market shifts, ‘no one’ is currently aiming to replace Nvidia GPUs fully, as the tech enterprise recognizes Nvidia’s superior capabilities in handling a variety of workloads.
Nvidia’s competitors are not idle either, as major tech players like Amazon, Microsoft, and OpenAI are exploring their custom processors to capitalize on market trends.
