Unlock Editor's Digest at no cost
FT editor Roula Khalaf selects her favourite stories on this weekly newsletter.
Nvidia made a surprise announcement on Sunday concerning the next generation of its AI processors, lower than three months after its most up-to-date launch.
At the Computex conference in Taipei, the chipmaker's CEO, Jensen Huang, unveiled “Rubin,” the successor to its “Blackwell” chips for data centers, that are currently in production after their announcement in March.
The unexpected move to unveil the following wave of products before Blackwell has even began shipping to customers shows how the world's most respected chipmaker is attempting to cement its dominance in AI processors, which has catapulted it into the ranks of the world's most respected corporations.
“A brand new era of computing is starting,” Huang said as Nvidia also announced latest AI chip deals with PC makers.
Rubin is scheduled to be delivered from 2026 and guarantees improved energy efficiency. The Silicon Valley-based company desires to allay concerns that the expansion of AI data centers by major technology corporations will put a strain on the ability grid in some regions.
The announcement contained few details, but Huang said Nvidia is working on constructing latest AI platforms on a “one-year cycle.”
Nvidia's pace of innovation has change into disproportionately vital to the general stock market as traders bet on whether the large AI-driven surge of a handful of huge technology corporations can proceed.
The chipmaker increased its market capitalization by about $350 billion after reporting a rapid increase in sales, moving the corporate onto Apple's side and becoming the second most respected U.S. company after Microsoft.
While Nvidia today sells the vast majority of the AI chips needed to coach large language models like OpenAI's GPT, the corporate faces growing competition from AMD and Intel, in addition to custom chips developed by cloud computing providers like Amazon, Google and Microsoft.
Nvidia’s Blackwell chip can be launched almost a 12 months after the presentation of the present “Hopper” chips.
The company also announced a brand new central processing unit based on Vera Arm on Sunday as Nvidia looks to supply more chips utilized in AI data centers.
CPUs have traditionally been the workhorses of any computer, a market dominated by Intel and AMD. But Huang is attempting to reshape the server market around his AI chips as artificial intelligence takes up a growing share of knowledge center workloads.
Nvidia began making graphics processors greater than 30 years ago to enrich Intel CPUs in video game PCs, but greater than 15 years ago, Huang realized the technology in his GPUs was also suitable for other data-intensive computing tasks, resembling AI.
The company is now attempting to boost its PC chip business by leveraging its dominant position in AI chips for data centers.
Huang also announced on Sunday deals with two PC makers, Asus and MSI, that may launch machines with Nvidia's GeForce RTX graphics processors to support a spread of AI tasks, from running digital assistants to video editing and encoding.
“Your future laptop can be continually helping you within the background,” Huang said. “The PC will run apps which can be enhanced by AI, from writing to photo editing to digital humans which can be AIs,” Huang said.
Nvidia has not specified when the Asus and MSI laptops will go on sale.
Various PC manufacturers and component suppliers are expected to make use of Computex to make announcements to position themselves as beneficiaries of an expected “AI PC” wave.
Microsoft recently unveiled a spread of AI-powered PCs and tablets equipped with its Copilot assistant tool and based on Qualcomm chips. The rollout is about to start later this month. Microsoft has announced that it plans to integrate Nvidia chips and AMD's Radeon graphics chips into its PCs in the long run.
PC sales have plummeted because the pandemic, but analysts expect that as demand rebounds, corporations will increasingly select AI PCs with powerful chips to run AI applications quite than relying solely on the cloud.
“AI PCs will bring essentially the most exciting innovation to the PC industry within the last two to 3 many years because the invention of the World Wide Web within the late Nineteen Eighties,” Morgan Stanley analysts wrote in a note last month.
They said running AI applications on devices is cheaper and more flexible than within the cloud, and in addition brings privacy advantages. AI PCs will account for about 65 percent of all PC shipments by 2028, up from 2 percent this 12 months, Morgan Stanley predicted.
