Nvidia CEO Jensen Huang has fired back at recent skepticism surrounding the company’s dominance in the AI chip market, directly addressing claims that competitors like China’s DeepSeek have achieved comparable AI capabilities with significantly less hardware. At a big event for computer programmers, Huang said that everyone has been wrong about how much computer power you need for AI.
“Almost the entire world got it wrong,” Huang said on stage at the Nvidia GPU Technology Conference (GTC) he called “the Super Bowl of AI”, as per news agency Reuters.
“The amount of computation we need as a result of agentic AI, as a result of reasoning, is easily 100 times more than we thought we needed this time last year,” he said, referring to AI agents that require little human intervention for routine tasks.
Huang emphasized that Nvidia is well-positioned to meet these changing demands. He highlighted the importance of both speed and scale in AI inference, stating, “If you take too long to answer a question, the customer is not going to come back. This is like web search.”
Nvidia launches next-generation GPU chip, Blackwell Ultra
Nvidia also announced a new generation of chips, including the next-generation GPU chip, Blackwell Ultra. Blackwell Ultra is set for release in the second half of this year and will feature increased memory compared to the current Blackwell chip, enabling the support of larger AI models.
The successor to Blackwell, the Vera Rubin chip system is slated for release in the second half of 2026 and will offer even faster processing speeds. Nvidia also plans to continue its chip development trajectory with the Feynman chips, expected to arrive in 2028.
These new chips are designed to optimise AI systems for both rapid and high-volume responses. Huang argued that Nvidia’s chips are uniquely capable of delivering both, a crucial factor for user satisfaction and retention.
