Nvidia and AMD Chiefs face off in a fight to take control of AI

(Bloomberg) — Nvidia Corp. bosses and Advanced Micro Devices Inc. unveiled new generations of chips that drive a global boom in AI development, deepening a rivalry that could decide the direction of artificial intelligence design and adoption.

Bloomberg’s Most Read

Jensen Huang and Lisa Su — both Taiwanese natives and now local celebrities from top U.S. tech powerhouses — employed different approaches to imparting their knowledge during back-to-back shows at the world’s biggest computing conference this week in Taipei.

Nvidia’s CEO has repeatedly expressed his $2.7 trillion company’s dominance in the accelerators that OpenAI and Microsoft Corp. are part of. rely on to build generative AI services like ChatGPT. Huang went so far as to tease a chip due in 2026, which he dubbed Rubin – after Vera Rubin, the American woman who helped discover dark matter. The chip, which will succeed the Blackwell family, will be fundamental to sustaining its unbridled leadership.

While Huang headlined his two-hour presentation on Sunday, AMD’s Su opted to make it more of a team effort. It brought in a range of high-profile partners, from HP Inc. CEO Enrique Lores to Lenovo Group Ltd.’s Luca Rossi, to convey the company’s focus on designing neural processors – a type of chip that performs AI directly from laptops. At one point during his speech at Computex, Asustek Computer Inc. President Jonney Shih called it “the pride of Taiwan” — a characterization often associated with Nvidia’s Huang, whose market value is now about 10 times greater than that of AMD.

“People see Nvidia as an embodiment of Jensen. And while Lisa is AMD’s savior, she makes it very clear that this is about everyone around her,” said Ian Cutress, chief analyst at consultancy More Than Moore. “AMD still has that underdog element to the business, and with AI that’s true.”

The back-to-back presentations – attended by hundreds of people and viewed by thousands of people around the world – underscore the growing interest in a technology that has the potential to redefine a range of industries and create new devices that can generate videos and other content almost instantly. content from simple commands.

There is a personal component to the competition as well. Su and Huang are not just Taiwanese, they were born in the same city of around 1.8 million people on the southern coast of Taiwan and are distantly related. This did not make either of them any more willing to give ground to the other.

The story continues

Su was originally chosen by organizers to kick off the week-long conference, one of the most important summits for technology executives around the world. Instead, Nvidia’s CEO hosted his own presentation on Sunday night, before the official opening of the meeting, to talk about the chipmaker’s strategic plans.

Huang, who had been the Computex speaker a year earlier, spoke at National Taiwan University just hours before the conference opened, essentially stealing the spotlight from his closest competitor. AMD’s Su on Monday obliquely referred to the episode during an OpenAI demo of a chatbot.

“Let’s start by telling the tool that we are interested in Taiwan and that we will participate in Computex,” said Su. “And you can ask something about the opening talk,” she joked, earning laughter from the audience.

Nvidia sees the rise of generative AI as a new industrial revolution and hopes to play an important role as the technology shifts to personal computers, the CEO said in his speech at National Taiwan University. He returned to themes he presented a year ago at the same venue, including the idea that those without AI capabilities will be left behind.

Computex delegates noted that Huang’s performance led to Nvidia’s continued dominance – a position that seemed difficult for AMD or any other rival to shake in the short term. A fund manager told Bloomberg News that Huang has generated a lot of buzz, especially around Rubin, although the CEO did not elaborate on a chip due in 2026.

Shares of Taiwan Semiconductor Manufacturing Co. and other suppliers rose after the announcement. TSMC shares rose 3.1%, while Wistron Corp. rose 3.1%. gained 1.8%.

Nvidia is selling customers a completely proprietary system where companies can buy their chips, networking equipment and everything else needed to run advanced AI development in data centers called ‘AI factories’. AMD, on the other hand, touts open standards that make its hardware interoperable with rivals like Intel Corp.

Huang said the upcoming Rubin AI platform will use HBM4, the next iteration of high-bandwidth essential memory that has become a bottleneck for AI accelerator production. Leading manufacturer SK Hynix Inc. is all but sold out until 2025. Otherwise, it hasn’t offered detailed specs for Nvidia’s upcoming products.

“Teasing Rubin and Rubin Ultra was extremely smart and is indicative of their commitment to a year-over-year upgrade cycle,” said Dan Newman, CEO and chief analyst at Futurum Group. “What I feel he has made clear most clearly is the cadence of innovation and the company’s relentless pursuit of maximizing the edge of technology, including software, process, packaging and partnerships to protect and expand its moat and market position.”

Read more: Samsung replaces Chip Chief after SK Hynix takes AI lead

But it was Su who commanded a broader, star-studded supporting lineup. She was joined on stage by Microsoft Windows chief Pavan Davuluri, a key leader at a company that is seen as having an edge over Google and Baidu Inc.

It and its partners used the Computex slot to showcase new AI-enhanced laptops from the Copilot+ brand. Most devices hitting the market are based on NPUs, which are more efficient in handling AI tasks and thus help extend battery life. Qualcomm Inc. was one of the first to launch Copilot+ PCs, but AMD’s response will be on the market starting in July. Nvidia has no comparable offering.

Su concluded his talk by returning to a key sphere of competition: data centers and AI training. She matched Huang’s annual upgrade schedule by announcing new AI chips for 2025 and 2026 and promising to accelerate the rate of improvement of her company’s semiconductors.

“For me, the big difference between the two is that Jensen is not just a technician and visionary, he is also a salesman. Lisa is much more focused on execution,” said Dave Altavilla, principal analyst at HotTech Vision and Analysis. “She has her work cut out for her and they are picking the locations where it matters, and the data center is obviously where the high margin is.”

–With assistance from Ian King and Nick Turner.

Bloomberg Businessweek Most Read

©2024 Bloomberg LP