Jensen Huang and Lisa Su have a lot in common. In addition to their respective positions as CEOs of chipmakers Nvidia (NVDA) and AMD (AMD), the two are both first-generation Americans hailing from the southern Taiwanese city of Tainan and are even distant cousins. As industry leaders in semiconductor manufacturing, they have also in recent years become key players amid the artificial intelligence (A.I.) boom. Huang and Su laid out their company roadmaps for the next generations of A.I. chips while taking the stage at Computex 2024, an annual tech trade show held in Taipei, Taiwan. Nvidia and AMD made a name for themselves with graphics processing units (GPUs) powering data centers that run generative A.I. models like OpenAI’s GPT and Google’s Gemini.
Yesterday (June 2), before the conference officially kicked off, Huang announced a new A.I. chip platform called “Rubin,” expected to roll out in 2026. The announcement came less than three months after Nvidia unveiled its next-generation A.I. chip Blackwell, which has yet to hit the market. “I’m not sure whether I’m going to regret this or not. We have code names in our company and we try to keep them very secret—oftentimes most of our employees don’t even know,” Huang said. Rubin is named after the U.S. astronomer Vera Rubin.
Both Blackwell and Rubin are in full development, said Huang, who noted they will be produced on a “one-year rhythm.” Blackwell will be made available later this year alongside the Blackwell Ultra in 2025 and the Rubin Ultra in 2027.
“The pace of product releases from Nvidia is jaw-dropping, not just because the products are so incredible but also because they’re launching or announcing new products every six months when it used to be that the standard was 12 to 18 months,” Cory Johnson, chief market strategist at Futurum Group, told Observer. “Everyone else is playing catch-up, including AMD.”
How AMD plans to catch up with Nvidia
Huang gave his presentation solo, joined only by a group of robots as he discussed his vision for “physical A.I.” as the next wave of the technology—one that will see A.I.-powered robots able to work among humans. During her keynote today, Su brought out a series of AMD partners including Microsoft Windows chief Pavan Davuluri and HP CEO Enrique Lores before she divulged details on AMD’s A.I. chip timeline.
Like Nvidia, AMD plans to develop new A.I. processors on an “annual cadence.” Following the launch of MI300X last year, the company in the fourth quarter of 2024 will make available its successor MI325X, which Su described as faster and offering more memory. This will be followed by the MI350 in 2025 and MI400 series in 2026. “It’s just so clear that the demand for A.I. is accelerating so much going forward,” said Su. “We’re really just at the beginning of a decade-long megacycle for A.I.”
Nvidia, which accounts for around 70 percent of A.I. semiconductor sales, has a market capitalization of $2.8 trillion, while AMD’s measures at around $264 billion. Nvidia’s success has made Huang the world’s 14th wealthiest person with an estimated net worth of $99.8 billion. Su, meanwhile, has for five consecutive years been ranked the highest-paid female CEO in the U.S.
Their recent announcements indicate a turning point in the tech industry, according to Johnson. “The pace of innovation is faster, and specifically the pace of product releases,” he said, adding that the developments are all the more impressive coming from Huang and Su. “It’s a pretty amazing thing, this day in the history of the world, to look up and see two Americans who were born in Taiwan leading innovation, really changing the world—and back in Taiwan talking about it.”
Alexandra Tremayne-Pengelly
Source link
