On Monday, Nvidia chief executive Jensen Huang unveiled the company's newest AI chip, the new Blackwell B200 GPU and GB200 "superchip."
Although Apple, Microsoft, and Amazon have also developed chips with AI in mind, they cannot fulfill their AI promises since they cannot obtain Nvidia's highly sought-after products.
Nvidia Reveals Blackwell B200
CEO Huang said on Monday at a developer's conference in California that graphics processors are vital in creating generative AI. "We need bigger GPUs. So, ladies and gentlemen, I would like to introduce you to a very, very big GPU."
Wedbush analyst Dan Ives has dubbed the event "AI Woodstock." It has become a must-attend event on the big tech calendar due to Nvidia's unique contribution to the AI revolution that has swept the globe since ChatGPT's launch in late 2022.
Huang joked as he took the stage in a packed venue more often used for concerts and ice hockey games. He told the audience he hoped they realized this was not a concert but a developer's conference.
Although competitors like AMD and Intel are still finding it difficult to match the power and efficiency of Nvidia's blockbuster H100 product, which was introduced in 2022, the company's strong GPU chips and software are crucial in developing generative AI.
Nvidia's share price increased by almost 25% in the previous year because of its pivotal role in the AI revolution. This puts it ahead of Amazon in terms of market value and behind only Microsoft and Apple.
Not letting up, Nvidia announced to the assembled developers and tech leaders that it would be launching a CPU with even more power and related software on a platform dubbed Blackwell, which was named for David Blackwell, the first Black scientist to be admitted into the National Academy of Science.
Furthermore, Nvidia claimed that all GPUs were AI "superchips" that could train AI models four times faster than the previous generation. Huang said that the rate at which computing is advancing is insane.
Nvidia also promised to deliver 25 times more energy efficiency, a crucial claim given that AI development is criticized for its voracious demand for natural resources and energy compared to more conventional computing. It reduces cost and energy consumption by up to 25 times over an H100.
Unlike its competitors, Intel, Micron, and Texas Instruments, Nvidia, like AMD, relies on subcontractors, mostly Taiwan Semiconductor Manufacturing Co., to make its chips.
Given the geopolitical tensions between Taiwan and China, this might be a weak point, and the US has prohibited Nvidia from selling its most powerful processors to Chinese companies.
Nvidia Announces Powered Robots
Other AI advancements from Nvidia, such as a platform for humanoid robot training, were also revealed.
Nvidia said that Project Gr00t was not named after the Groot character from the Guardians of the Galaxy film. However, it was called the "world's first human foundation model."
According to Nvidia, Gr00t-powered robots will be made to mimic human speech and gestures and learn how to communicate with the outside world through experience.
It claimed that the models would enable a robot to learn from a handful of human demonstrations to help with everyday tasks and emulate human movement by observing them.
Furthermore, it also collaborated with Apple to include AI features in the newly launched Vision Pro spatial computing equipment.
The collaboration responds to Apple's mounting pressure to demonstrate that it can keep up with Google, Amazon, Meta, and OpenAI in AI.
Nvidia also revealed the Earth-2 Cloud Platform, which uses AI supercomputer modeling to forecast climate change.
Related Article : Nvidia Boss Predicts Death of Coding as Kids Rely on AI
