GTC24: Blackwell GPUs, quantum cloud, robotic AI and everything else NVIDIA's Jensen Huang unveiled

NVIDIA chief Jensen Huang at GTC24

One moment the stadium was lit up in pink, and in the next it changed to blue. While some boasted of their previous pilgrimages to the event, it was a maiden voyage for others. The occasion could easily have been mistaken for a music concert.

At a packed hockey arena in San Jose, a city in the lap of Silicon Valley, around 12,000 developers, startup founders, venture capitalists and big tech executives sat in rapt attention as the chief of one of the most important corporations in the age of artificial intelligence spoke for a straight two hours at NVIDIA's GTC AI conference 2024.

For the past five years, NVIDIA's annual GPU (graphics processing unit) technology conference has been a virtual affair. A lot has changed in the interim, from the shock of the Covid pandemic, which catalysed digital adoption, to the dawn of generative AI.

From a market capitalisation of $144 billion at the end of 2019, the tech giant's stock has jumped 15-fold to make it the third-most-valuable company in the world, with a market value of over $2.2 trillion.

Much of this growth has been a result of its dominance in making semiconductor chips for generative AI applications such as ChatGPT and Gemini.

But growth is a beast that always asks for more. As is quite often expected of a leader in any field, the tech ecosytem wanted to know what's next.

At Monday's keynote address in Silicon Valley, Huang had his job cut out: To impress upon the tech world that NVIDIA is still galloping ahead of the competition in the AI race.

Here are all the new products and projects that he announced today:

Blackwell GPUs

In line with expectations, Huang announced the next big leap for computing power in AI. While the industry is still dealing with a supply constraint of NVIDIA's previous two generations of AI chips, the company is ready with the next version, which is 30 times more powerful than Hopper, its immediate predecessor.

"This is Hopper. Hopper changed the world. This is Blackewell. It's okay, Hopper. You're good. Good girl," Huang said while holding the two chips on the stage.

AI heft for Apple Vision Pro

Omniverse, an Nvidia tool that allows developers to simulate physical experiences digitally, will now be available to developers building content for Apple's Vision Pro headset.

In a demo unveiled at the global AI conference, NVIDIA presented an interactive, physically accurate digital twin of a car streamed in full fidelity to Apple Vision Pro’s high-resolution displays.

The demo featured a designer wearing the Vision Pro. The designer toggles through paint and trim options and even enters the vehicle — as Omniverse's spatial computing blends 3D photorealistic environments with the physical world.

Becoming the TSMC of AI

NVIDIA launched a service that businesses can use to create and deploy their own AI applications, while retaining full ownership and control of their intellectual property.

Announcing the service, Huang said that NVIDIA wants to be an AI foundry similar to the way Taiwanese chip fab company TSMC is the world's most important semiconductor foundry.

“Established enterprise platforms are sitting on a goldmine of data that can be transformed into generative AI copilots,” said Huang. “Created with our partner ecosystem, these containerised AI microservices are the building blocks for enterprises in every industry to become AI companies," he added.

The chip is named in honor of David Harold Blackwell, a mathematician who specialised in game theory and statistics, and the first Black scholar inducted into the National Academy of Sciences. The new architecture succeeds the NVIDIA Hopper architecture launched two years ago.

An AI platform for humanoid robots

NVIDIA also announced Project GR00T, a general-purpose foundation model for humanoid robots.

Robots built on GR00T will be designed to understand natural language and emulate movements by observing human actions—quickly learning coordination, dexterity, and other skills in order to navigate, adapt, and interact with the real world. In his GTC keynote, Huang demonstrated several such robots completing a variety of tasks.

"Building foundation models for general humanoid robots is one of the most exciting problems to solve in AI today,” said the founder and CEO of NVIDIA. “The enabling technologies are coming together for leading roboticists around the world to take giant leaps towards artificial general robotics.”

Quantum cloud

NVIDIA also launched a simulation platform on the cloud that can be used by scientists for quantum computing and algorithm research.

This will allow researchers to generate large, sophisticated quantum programs, as well as deeply analyse and execute quantum circuits. Further, it will help scientists tackle complex quantum chemistry problems, such as molecular simulation.

AI to forecast weather

NVIDIA today announced its Earth-2 climate digital twin cloud platform to simulate and visualise weather and climate.

When combined with proprietary data owned by companies in the $20 billion climate tech industry, the Earth-2 application programming interfaces help users deliver warnings and updated forecasts in seconds, compared to the minutes or hours in traditional CPU-driven modeling.

"Climate disasters are now normal—historic droughts, catastrophic hurricanes and generational floods appear in the news with alarming frequency,” said Huang.

“Earth-2 cloud APIs strive to help us better prepare for—and inspire us to act to moderate—extreme weather," he added.

(This correspondent was in San Jose at the invitation of Nvidia)

Event alert: Moneycontrol and CNBC TV18 are hosting the ultimate event on artificial intelligence, bringing together entrepreneurs, ecosystem enablers, policymakers, industry leaders, and innovators on March 27 in Pune. Click here to register and gain access to the AI Alliance Pune Chapter.