Taipei/Hong Kong
Nvidia, AMD and Intel have individually launched the next generation of their synthetic intelligence (AI) chips in Taiwan, as a three-way race heats up.
Jensen Huang, CEO of Nvidia (NVDA), mentioned Sunday that the corporate would roll out its most superior AI chip platform, known as Rubin, in 2026.
The Rubin platform will succeed the Blackwell, which provides chips for information facilities and was introduced solely in March. It was dubbed by Nvidia at the time because the “world’s most powerful chip.”
The Rubin will characteristic new graphics processing items (GPUs), a brand new central processing unit (CPU) known as Vera and superior networking chips, Huang mentioned in an handle at National Taiwan University in Taipei.
“Today, we’re at the cusp of a major shift in computing,” Huang advised the viewers forward of the opening of Computex, a tech commerce present organized yearly in Taiwan. “The intersection of AI and accelerated computing is set to redefine the future.”
He revealed a roadmap for brand spanking new semiconductors that can arrive on a “one-year rhythm.”
Investors have been driving up shares in chip companies driving the generative AI growth. Shares of Nvidia, the market chief, have greater than doubled over the previous 12 months.
“Nvidia clearly intends to keep its dominance for as long as possible and in the current generation, there is nothing really on the horizon to challenge that,” mentioned Richard Windsor, founding father of Radio Free Mobile, a analysis firm specializing in the digital and cellular ecosystem.
Nvidia accounts for round 70% of AI semiconductor gross sales. But competitors is rising, with main rivals AMD (AMD) and Intel (INTC) introducing new merchandise in an effort to problem Nvidia’s dominance.
On Monday, AMD CEO Lisa Su unveiled the corporate’s newest AI processors in Taipei and a plan to develop new merchandise over the next two years.
Its next generation MI325X accelerator will probably be made out there in the fourth quarter of this 12 months, she mentioned.
A day later, Intel CEO Patrick Gelsinger introduced the sixth generation of its Xeon chips for information facilities and its Gaudi 3 AI accelerator chips. He touted the latter, which competes with Nvidia’s H100, as being one-third cheaper than its rivals.
The world competitors to create generative AI purposes has led to hovering demand for the leading edge chips used in information facilities to assist these packages.
Both Nvidia and AMD, that are run by Taiwan-born American CEOs who’re a part of the identical household, have been as soon as greatest identified by players for promoting GPUs that show visuals in video video games, serving to them come to life.
While the 2 nonetheless compete in that house, their GPUs at the moment are additionally being used to energy generative AI, the technology that underpins newly well-liked methods equivalent to ChatGPT.
“AI is our number one priority, and we’re at the beginning of an incredibly exciting time for the industry,” Su added.
“We launched MI300X last year with leadership inference performance, memory size and compute capabilities, and we have now expanded our roadmap so it’s now on an annual cadence, that means a new product family every year,” she mentioned.
The new chip will succeed the MI300 and have extra reminiscence, quicker reminiscence bandwidth and higher laptop efficiency, Su added. The firm will launch a brand new product household yearly, with the MI350 set for 2025 and MI400 a 12 months later.