The Nvidia GTC that was held on 16 March, Monday, saw the chipmaking giant make several key announcements, with CEO Jensen Huang saying that the company could generate as much as $1 trillion in revenue by 2027 in sales of AI chips alone, amid the ongoing artificial intelligence bloom.
Dressed in a black leather jacket like always, Nvidia CEO Jensen Huang spoke for nearly three hours at the event. During his speech, Huang highlighted how his company was building the hardware, software and infrastructure it needs to dominate the AI industry.
Here are the key things that Nvidia and Jensen Huang announced during the company’s GTC meet.
$1 trillion revenue
One of the key highlights in Jensen Huang’s speech was the potential of Nvidia generating $1 trillion revenue through 2027. Addressing thousands at a packed arena in San Jose, California, he said he expects purchase orders between Blackwell and Vera Rubin to reach $1 trillion through 2027.
“I believe that computing demand has increased by 1 million times in the last two years,” Huang said. “It is the feeling that we all have. It is the feeling every startup has.”
Nvidia Groq 3 Language Processing Unit
Speaking further on Monday, Jensen Huang also revealed the Nvidia Groq 3 Language Processing Unit, or LPU. This is the first chip from the startup that Nvidia mostly acquired in a $20 billion asset purchase in December last year. The LPU is likely to ship in the third quarter. It is built to Nvidia’s GPU processing technology, with the core optimised to speed it up.
Vera Rubin system
“We’re working with our partners on a new computer called Vera Rubin Space One,” Huang said. “It is going to go out to space and start data centres.”
Partners in the project include Starcloud, which is planning a November satellite launch that will mark the “cosmic debut” of the new Nvidia module.
Kyber prototype
During the Nvidia GTC on Monday, Jensen Huang also unveiled a prototype for Kyber, the company’s next big rack architecture jump following Rubin. The Kyber system is expected to integrate 144 GPUs in compute trays. They will be stacked vertically instead of horizontally to boost density and lower latency.
The Kyber design is expected to be available in Nvidia’s next rack-scale system Vera Rubin Ultra. It is likely to ship in 2027.
Automotive updates
Jensen Huang also shared updates on the automotive front. He gave details on the previously announced agreement with Uber, saying that the ride-hailing service is set to launch a fleet powered by Nvidia’s Drive AV software by 2028. This will be available across 28 cities across four continents starting with Los Angeles and San Francisco next year.
I believe that computing demand has increased by 1 million times in the last two years.
Huang further announced that Nissan, BYD, Geely, Isuzu and Hyundai are manufacturing level 4 autonomous vehicles on Nvidia’s Drive Hyperion program. Along with this, Isuzu and China-based Tier IV are also building autonomous buses using the platform.
Nvidia has accelerated its technology development in recent years, seeking to replace its entire product lineup on an annual basis while adding new components.
