Meta Platforms Inc is in talks to spend billions on Google’s artificial intelligence (AI) chips, The Information reported on Monday, adding to a monthslong share rally as the search giant has made the case it can rival Nvidia Corp as a leader in AI technology.
A deal would signal growing momentum for Google’s chips and long-term potential to challenge Nvidia’s market dominance, after the company earlier agreed to supply up to 1 million chips to Anthropic PBC.
Meta is in discussions to use the Google chips — known as tensor processing units (TPUs) — in data centers in 2027, The Information reported, citing an unidentified person familiar with the talks. Meta also may rent chips from Google’s cloud division next year, the news outlet said.
Photo: Reuters
An agreement would help establish TPUs as an alternative to Nvidia’s graphics processing units (GPUs), the gold standard for big tech firms and start-ups from Meta to OpenAI that need computing power to develop and run AI platforms.
Nvidia’s stock is already facing headwinds as investors fear a broader AI bubble. Michael Burry, immortalized in The Big Short for his bets against the housing market during the 2008 financial crisis, has scrutinized the chipmaker over circular AI deals, hardware depreciation and revenue recognition.
After Google’s Anthropic deal was announced, Seaport analyst Jay Goldberg called it a “really powerful 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.
“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,” a Google spokesperson said.
Representatives for Meta declined to comment.
“We’re delighted by Google’s success – they’ve made great advances in AI, and we continue to supply to Google,” an Nvidia spokesperson said on Tuesday, echoing a post the company put on the social media platform X. “Nvidia is a generation ahead of the industry – it’s the only platform that runs every AI model and does it everywhere computing is done. Nvidia offers greater performance, versatility, and fungibility than ASICs, which are designed for specific AI frameworks or functions.”
A deal with Meta — one of the biggest spenders globally on data centers and AI development — would mark a win for Google. But much depends on whether the tensor chips can demonstrate the power efficiency and computing muscle necessary to become a viable option in the long run.
The tensor chip — first developed more than 10 years ago especially for artificial intelligence tasks — is gaining momentum outside its home company as a way to train and run complex AI models. Its allure as an alternative has grown at a time companies around the world worry about an overreliance on Nvidia, in a market where even Advanced Micro Devices Inc is a distant runner-up.
GPUs, the part of the chip market dominated by Nvidia, were created to speed the rendering of graphics — mainly in video games and other visual-effects applications — but turned out to be well-suited to training AI models because they can handle large amounts of data and computations. TPUs, on the other hand, are a type of specialized product known as application-specific integrated circuits, or microchips that were designed for a discrete purpose.
The tensor chips were also adapted as an accelerator for AI and machine learning tasks in Google’s own applications. Because Google and its DeepMind unit develop cutting-edge AI models like Gemini, the company has been able to take lessons from those teams back to the chip designers. At the same time, the ability to customize the chips has benefited the AI teams.
CHIP RACE: Three years of overbroad export controls drove foreign competitors to pursue their own AI chips, and ‘cost US taxpayers billions of dollars,’ Nvidia said China has figured out the US strategy for allowing it to buy Nvidia Corp’s H200s and is rejecting the artificial intelligence (AI) chip in favor of domestically developed semiconductors, White House AI adviser David Sacks said, citing news reports. US President Donald Trump on Monday said that he would allow shipments of Nvidia’s H200 chips to China, part of an administration effort backed by Sacks to challenge Chinese tech champions such as Huawei Technologies Co (華為) by bringing US competition to their home market. On Friday, Sacks signaled that he was uncertain about whether that approach would work. “They’re rejecting our chips,” Sacks
NATIONAL SECURITY: Intel’s testing of ACM tools despite US government control ‘highlights egregious gaps in US technology protection policies,’ a former official said Chipmaker Intel Corp has tested chipmaking tools this year from a toolmaker with deep roots in China and two overseas units that were targeted by US sanctions, according to two sources with direct knowledge of the matter. Intel, which fended off calls for its CEO’s resignation from US President Donald Trump in August over his alleged ties to China, got the tools from ACM Research Inc, a Fremont, California-based producer of chipmaking equipment. Two of ACM’s units, based in Shanghai and South Korea, were among a number of firms barred last year from receiving US technology over claims they have
BARRIERS: Gudeng’s chairman said it was unlikely that the US could replicate Taiwan’s science parks in Arizona, given its strict immigration policies and cultural differences Gudeng Precision Industrial Co (家登), which supplies wafer pods to the world’s major semiconductor firms, yesterday said it is in no rush to set up production in the US due to high costs. The company supplies its customers through a warehouse in Arizona jointly operated by TSS Holdings Ltd (德鑫控股), a joint holding of Gudeng and 17 Taiwanese firms in the semiconductor supply chain, including specialty plastic compounds producer Nytex Composites Co (耐特) and automated material handling system supplier Symtek Automation Asia Co (迅得). While the company has long been exploring the feasibility of setting up production in the US to address
OPTION: Uber said it could provide higher pay for batch trips, if incentives for batching is not removed entirely, as the latter would force it to pass on the costs to consumers Uber Technologies Inc yesterday warned that proposed restrictions on batching orders and minimum wages could prompt a NT$20 delivery fee increase in Taiwan, as lower efficiency would drive up costs. Uber CEO Dara Khosrowshahi made the remarks yesterday during his visit to Taiwan. He is on a multileg trip to the region, which includes stops in South Korea and Japan. His visit coincided the release last month of the Ministry of Labor’s draft bill on the delivery sector, which aims to safeguard delivery workers’ rights and improve their welfare. The ministry set the minimum pay for local food delivery drivers at
RSS