Nvidia Corp is building a new business unit focused on designing bespoke chips for cloud computing firms and others, including advanced artificial intelligence (AI) processors, nine sources familiar with its plans told Reuters.
The dominant global designer and supplier of AI chips aims to capture a portion of an exploding market for custom AI chips and shield itself from the growing number of companies pursuing alternatives to its products.
The Santa Clara, California-based company controls about 80 percent of the high-end AI chip market, a position that has sent its stock market value up 40 percent so far this year to US$1.73 trillion after it more than tripled last year.
Photo: Reuters
Nvidia’s customers, which include ChatGPT creator OpenAI, Microsoft Corp, Alphabet Inc and Meta Platforms Inc, have raced to snap up the dwindling supply of its chips to compete in the fast-emerging generative AI sector.
The firm’s H100 and A100 chips serve as a generalized, all-purpose AI processor for many of those major customers, but the tech companies have started to develop their own internal chips for specific needs. Doing so helps reduce energy consumption, and can potentially shrink the cost and time to design.
Nvidia is now attempting to play a role in helping these companies develop custom AI chips that have flowed to rival firms such as Broadcom Inc and Marvell Technology Group Inc, the sources said.
“If you’re really trying to optimize on things like power, or optimize on cost for your application, you can’t afford to go drop an H100 or A100 in there,” Eclipse Ventures LLC general partner Greg Reichow said in an interview. “You want to have the exact right mixture of compute and just the kind of compute that you need.”
Nvidia does not disclose H100 prices, which are higher than for the prior-generation A100, but each chip can sell for US$16,000 to US$100,000 depending on volume and other factors. Meta plans to bring its total stock to 350,000 H100s this year.
Nvidia officials have met with representatives from Amazon.com, Meta, Microsoft, Google and OpenAI to discuss making custom chips for them, two sources familiar with the meetings said. Beyond data center chips, Nvidia has pursued telecom, automotive and video game customers.
Dina McKinney, a former Advanced Micro Devices Inc and Marvell executive, heads Nvidia’s custom unit and her team’s goal is to make its technology available for customers in cloud, 5G wireless, video games and automotives, a LinkedIn profile said. Those mentions were scrubbed and her title changed after Reuters sought comment from Nvidia.
According to estimates from research firm 650 Group’s Alan Weckel, the data center custom chip market will grow to as much as US$10 billion this year, and double that next year.
The broader custom chip market was worth roughly US$30 billion last year, which amounts to roughly 5 percent of annual global chip sales, according to Needham & Co analyst Charles Shi.
Currently, custom silicon design for data centers is dominated by Broadcom and Marvell.
In a typical arrangement, a design partner such as Nvidia would offer intellectual property and technology, but leave the chip fabrication, packaging and additional steps to Taiwan Semiconductor Manufacturing Co (台積電) or another contract chip manufacturer.
Nvidia moving into this territory has the potential to eat into Broadcom and Marvell sales.
“With Broadcom’s custom silicon business touching US$10 billion, and Marvell’s around US$2 billion, this is a real threat,” SemiAnalysis founder Dylan Patel said. “It’s a real big negative — there’s more competition entering the fray.”
Nvidia is in talks with telecom infrastructure builder Ericsson AB for a wireless chip that includes the chip designer's graphics processing unit technology, two sources familiar with the discussions said.
Weckle expects the telecom custom chip market to remain flat at roughly US$4 billion to US$5 billion a year.
Nvidia also plans to target the automotive and video game markets, according to sources and public social media postings.
Weckel expects the custom auto market to grow consistently from its current US$6 billion to US$8 billion range at 20 percent a year, and the US$7 billion to US$8 billion video game custom chip market could increase with the next-generation consoles from Microsoft and Sony Group Corp.
Nintendo Co's current Switch handheld console already includes Nvidia's Tegra X1 chip. A new version of the Switch console expected this year is likely to include a Nvidia custom design, one source said.
Hon Hai Precision Industry Co (鴻海精密) yesterday said that its research institute has launched its first advanced artificial intelligence (AI) large language model (LLM) using traditional Chinese, with technology assistance from Nvidia Corp. Hon Hai, also known as Foxconn Technology Group (富士康科技集團), said the LLM, FoxBrain, is expected to improve its data analysis capabilities for smart manufacturing, and electric vehicle and smart city development. An LLM is a type of AI trained on vast amounts of text data and uses deep learning techniques, particularly neural networks, to process and generate language. They are essential for building and improving AI-powered servers. Nvidia provided assistance
STILL HOPEFUL: Delayed payment of NT$5.35 billion from an Indian server client sent its earnings plunging last year, but the firm expects a gradual pickup ahead Asustek Computer Inc (華碩), the world’s No. 5 PC vendor, yesterday reported an 87 percent slump in net profit for last year, dragged by a massive overdue payment from an Indian cloud service provider. The Indian customer has delayed payment totaling NT$5.35 billion (US$162.7 million), Asustek chief financial officer Nick Wu (吳長榮) told an online earnings conference. Asustek shipped servers to India between April and June last year. The customer told Asustek that it is launching multiple fundraising projects and expected to repay the debt in the short term, Wu said. The Indian customer accounted for less than 10 percent to Asustek’s
‘DECENT RESULTS’: The company said it is confident thanks to an improving world economy and uptakes in new wireless and AI technologies, despite US uncertainty Pegatron Corp (和碩) yesterday said it plans to build a new server manufacturing factory in the US this year to address US President Donald Trump’s new tariff policy. That would be the second server production base for Pegatron in addition to the existing facilities in Taoyuan, the iPhone assembler said. Servers are one of the new businesses Pegatron has explored in recent years to develop a more balanced product lineup. “We aim to provide our services from a location in the vicinity of our customers,” Pegatron president and chief executive officer Gary Cheng (鄭光治) told an online earnings conference yesterday. “We
LEAK SOURCE? There would be concern over the possibility of tech leaks if TSMC were to form a joint venture to operate Intel’s factories, an analyst said Taiwan Semiconductor Manufacturing Co (TSMC, 台積電) yesterday stayed mum after a report said that the chipmaker has pitched chip designers Nvidia Corp, Advanced Micro Devices Inc and Broadcom Inc about taking a stake in a joint venture to operate Intel Corp’s factories. Industry sources told the Central News Agency (CNA) that the possibility of TSMC proposing to operate Intel’s wafer fabs is low, as the Taiwanese chipmaker has always focused on its core business. There is also concern over possible technology leaks if TSMC were to form a joint venture to operate Intel’s factories, Concord Securities Co (康和證券) analyst Kerry Huang (黃志祺)
RSS