NVIDIA’s GTC 2025: Jensen Huang Declares AI’s “Super Bowl” as Stock Dips

By InvestorHire | March 2025

SAN JOSE, CA – With the energy of a rock concert and the stakes of a tech revolution, NVIDIA CEO Jensen Huang took center stage at the company’s annual GTC conference on Tuesday, ready to shake up the AI world. Dressed in his signature black leather jacket, Huang kicked off the event by launching t-shirts into the roaring crowd of 25,000 attendees at the SAP Center, calling it the "Super Bowl of AI."

“The only difference is everybody wins at this Super Bowl,” Huang joked, setting the tone for what would be a mix of breakthrough announcements, AI-powered showmanship, and a rollercoaster ride for NVIDIA’s stock price.

AI Goes Personal: The DGX Revolution

Among the most exciting reveals was NVIDIA’s DGX Spark, the world’s smallest AI supercomputer, designed to bring data-center-level power straight to researchers, developers, and students. Powered by the NVIDIA GB10 Grace Blackwell Superchip, Spark delivers an eye-watering 1,000 trillion operations per second—giving AI innovators a powerful desktop tool to build, fine-tune, and deploy next-gen models.

Meanwhile, NVIDIA DGX Station takes desktop AI even further with the GB300 Grace Blackwell Ultra Superchip—featuring 784GB of memory and an ultra-high-speed NVIDIA ConnectX-8 SuperNIC, making AI training faster and more seamless than ever before.

And the best part? You don’t need a data center to run it. DGX users can now prototype on their desktops and scale up to DGX Cloud with virtually no code changes. That means faster innovation and fewer headaches for developers—a move that could change the game for AI-native computing.

DeepSeek, Dynamo, and a 30X Speed Boost

Despite NVIDIA’s dominance, AI skeptics have kept an eye on China’s DeepSeek R1, a reasoning model that promises AI intelligence at a fraction of NVIDIA’s cost. Back in January, this competitor sent shockwaves through the industry, contributing to a $589 billion market cap drop for NVIDIA in a single day.

But Huang wasn’t about to let DeepSeek steal the show. Enter NVIDIA Dynamo, an open-source software platform designed to make DeepSeek R1 up to 30 times faster—without needing extra GPUs.

To prove his point, NVIDIA’s VP of Hyperscale and HPC, Ian Buck, played a demo video showing exactly how Dynamo optimizes NVIDIA’s Blackwell chips for reasoning models.

“Dynamo can capture that benefit and deliver 30 times more performance in the same number of GPUs,” Buck told the crowd.

Translation? NVIDIA just turned a major competitor’s selling point into an afterthought.

Stock Dips as Chip Timelines Announced

But not everything was a victory lap. While Huang’s address had the crowd fired up, Wall Street had other plans. As soon as NVIDIA announced its GPU rollout schedule, its stock dropped more than 3% during his speech.

The timeline includes:

• Blackwell Ultra (2025)

• Vera Rubin (2026)

• Feynman (2027)

Some investors had hoped for faster release cycles, leading to a brief selloff. But Huang remained unfazed, dismissing concerns about lower AI spending and instead doubling down on the long-term impact of NVIDIA’s advancements.

Final Thoughts: The AI Race is Just Getting Started

Despite the stock hiccup, NVIDIA’s GTC 2025 made one thing clear: the company isn’t just leading AI—it’s rewriting the rules of the game. From supercomputers on your desk to open-source breakthroughs, NVIDIA is making sure AI is accessible, scalable, and faster than ever before.

And if Huang has his way, AI’s "Super Bowl" is just getting started.