Nvidia CEO Jensen Huang tried to reassure investors while his company was at the heart of the artificial intelligence boom and unveiling its latest products, including the much-anticipated Blackwell Ultra Computer chip.
Huang said the need for computer power enabled by advanced AI computer chips has grown exponentially compared to a year ago as more companies are moving towards so-called “inference” models that allow more companies to perform more complex tasks than basic chatbots.
“The whole world was almost wrong,” fans said. He riffed on stage for two hours without a teleprompter at Nvidia's GTC conference, dubbed “Super Bowl AI.” While wrapped in his trademark leather jacket and jeans.
“As a result of inference, the amount of computation required as a result of Agent AI is easily 100 times more than I thought this time last year,” says Huang.
Huang's bullish remarks at the meeting failed to make landfall with investors. Nvidia's shares sank more than 3% after trading on Tuesday. The stock closed at $115.73, falling nearly 20% over the past month.
Wall Street is seeking clarity from Nvidia and other major tech players following a major sale that sparked by Chinese startup DeepSeek in January.
This release raised concerns that Nvidia's chips were less necessary than previously thought in AI development. So far, companies like Meta and Microsoft have said they are all ahead of AI-related spending.
During the event, Huang confirmed that the next GPU chip, the Blackwell Ultra, will begin shipping later this year. Nvidia bosses have sparked demand for “Amazing” products.
Another upgraded chip family, called Bella Rubin, is stronger than Blackwell and allows for faster processing speeds, but is on track for release in late 2026, Huang added.
“If it takes too long to answer a question, the customer won't come back. This is like web search,” Huang says.
Elsewhere, Huang revealed that Nvidia will partner with General Motors to help automakers develop self-driving vehicles.
NVIDIA technology will be used to support the manufacturing and design process, as well as in-vehicle computer systems, he said.
“The self-driving car time has arrived,” fans said.
With post wire
