The event drew more than 25,000 people to the SAP Center’s National Hockey League arena, and Huang opened the keynote by launching t-shirts into the crowd and coronating this year’s GTC the “Super Bowl of AI.”

“The only difference is everybody wins at this Super Bowl, everybody’s a winner,” he joked. And like the Super Bowl, there were GTC watch parties and packed crowds to get a glimpse of Huang on stage. 

With his address, Huang sought to dispel any uneasiness around AI investment, and said discussion about lower spending does not concern Nvidia. In January, apprehension engulfed the chip maker after it lost $589 billion in market cap in a single day after Chinese AI reasoning model Deepseek R1 claimed to operate at a fraction of the cost. 

While large language models offer foundational knowledge, reasoning models offer more complex, analytical responses. Using the company’s new open source software Nvidia Dynamo, Huang said the tech giant’s Blackwell chips will be able to make DeepSeek R1 30 times faster. He then played a video demonstrating for the crowd how it could be done.

“Dynamo can capture that benefit and deliver 30 times more performance in the same number of GPUs in the same architecture for reasoning models like DeepSeek,” said Ian Buck, vice president and general manager of Nvidia’s hyperscale and HPC computing business.

From there, Huang’s keynote covered everything from the chip maker’s plans to roll out its newest chips— Blackwell Ultra later this year, Vera Rubin in 2026, and Feynman in 2027.

“We have an annual rhythm of roadmaps that has been laid out for you,” Huang said.

While Nvidia’s announced its strategic runway for years to come, that wasn’t enough to stop the stock’s slide. The chip maker’s share price tumbled 3.4% Tuesday.