Nvidia CEO Jensen Huang introduces a ‘big, big GPU’ that is ‘pushing the limits of physics’

For his two-hour lengthy keynote deal with at Nvidia’s GTC builders convention yesterday on the packed SAP Heart in San Jose, CEO Jensen Huang was clad in a black leather-based jacket that was only a bit extra rock and roll than the plain variations he has sported over the previous few years. With a number of silver zippers right here, a little bit shine and texture there, the jacket supplied a small clue that, whereas Huang joked that he hoped the viewers realized the gathering was not a live performance, Nvidia is among the previous yr’s greatest AI rock stars.

However Huang didn’t tout the corporate’s inventory value, which has skyrocketed since OpenAI’s ChatGPT was launched in November 2022 and generative AI exploded — with its accompanying starvation for Nvidia’s GPUs. There was no want to say generative AI fashions require 1000’s of GPUs to run — and Nvidia holds over 80% of the GPU market, according to John Peddie Analysis. Or that OpenAI reportedly used 10,000 Nvidia GPUs to coach ChatGPT. Or that Meta CEO Mark Zuckerberg not too long ago mentioned the corporate plans to purchase 350,000 Nvidia H100 GPUs.

Huang mentioned Nvidia’s new Blackwell GPU affords 30x efficiency enhance

As an alternative, Huang sought to indicate off Nvidia’s dominance by unveiling the corporate’s latest AI chip — a “massive, massive GPU” known as Blackwell, which, he mentioned, affords a 30 occasions efficiency enhance for LLM inference workloads in comparison with earlier iterations and is “pushing the bounds of physics.” Just like the well-known scene in “Jaws” when Roy Scheider’s Chief Brody says “You’re going to wish an even bigger boat,” Huang defined that the ‘new industrial revolution’ of generative AI requires super-fast information facilities and greater GPUs to deal with the mammoth scale of generated tokens, that, in flip, create ‘extremely precious software program’ in every part from healthcare to robotics.

With a tech-heavy presentation that went deep into the maths behind the dimensions of at this time’s generative AI fashions, it was clear that Huang was talking on to the developer neighborhood that’s core to Nvidia’s success. “I need to present you the soul of Nvidia, the soul of our firm, on the intersection of pc graphics, physics and synthetic intelligence, all intersecting inside a pc,” he mentioned.

Huang didn’t deal with wider societal influence of AI

With the deal with constructing, accelerating, exploring unusual new worlds, boldly going the place nobody has gone earlier than — oh, wait, that’s Star Trek — it was additionally notable that Huang was not chatting with a wider public viewers that has many questions on AI that don’t should do with PFLOPS.

There have been no feedback concerning the societal impacts of AI — from impacts to the workforce (apart from making it extra environment friendly) and job abilities to problems with ethics and belief, bias and misinformation. Apart from many mentions of power ‘effectivity,’ there was no point out of how the manufacturing and use of Nvidia merchandise influence the surroundings.

As an alternative, after an look by Disney’s cute Star Wars AI robots, which realized to stroll in Nvidia’s robotics simulation platform Isaac Sim, Huang drilled down on his predicted “new industrial revolution” of accelerated computing — with a 100T generative AI trade powered by the Blackwell platform; Nvidia’s new NIM container microservices to simply distribute gen AI software program; its NEMO framework and AI foundry; and Omniverse robotics platform.

The keynote ended with a video displaying a small Star Trek-like plane piloted by an animated Jensen Huang reaching for the celebrities and “powering the brand new period of computing.” Nvidia’s builders will definitely be on board. The query is, is the remainder of society prepared for what could be a tough trip?