This week, Nvidia CEO Jensen Huang will take the stage at the company’s annual developer conference, aiming to solidify its position as the leader in the AI chip market. With a valuation nearing $3 trillion, Nvidia faces growing pressure as its biggest customers seek to cut soaring AI costs.

The conference comes on the heels of a shakeup in the AI industry. China’s DeepSeek recently unveiled a chatbot that reportedly requires less computing power than rivals, spooking U.S. markets and causing Nvidia’s stock to dip. Nvidia’s revenue, which has surged over 300% in three years to $130.5 billion, relies heavily on selling high-powered chips that cost tens of thousands of dollars each.

New Chip on the Horizon

At the event, Nvidia is expected to unveil details about its new Vera Rubin chip system, named after the astronomer who discovered dark matter. The system is slated for mass production later this year. This follows the delayed rollout of its predecessor, the Blackwell chip, which has already started hitting the market.

Shifting AI Markets

Nvidia’s dominance is being tested as the AI industry shifts focus. While the company controls over 90% of the “training” market—where AI models are fed massive datasets to learn—it faces stiff competition in “inference,” where trained models generate answers for users.

Startups and rivals like AMD are betting on cheaper, more energy-efficient chips for inference tasks. Nvidia’s power-hungry chips have even led some AI firms to explore nuclear energy as a power source.

“They have a hammer, and they’re just making bigger hammers,” said Bob Beachler of Untether AI, one of many startups challenging Nvidia in inference.

The Rise of Reasoning AI

Nvidia, however, is banking on a new AI trend called “reasoning.” These chatbots think aloud, generating and analyzing text in real-time—a process that demands the heavy computing power Nvidia’s chips excel at.

“The inference market will dwarf the training market,” said Jay Goldberg of D2D Advisory. “While Nvidia’s market share may shrink, the overall revenue potential could be enormous.”

Beyond Chatbots: Robots and Quantum Computing

Nvidia is also expected to showcase its ambitions beyond AI chatbots. The company plans to use AI to enhance robotics and delve into quantum computing. Earlier this year, Huang’s comments downplaying quantum computing’s near-term potential sparked controversy, prompting Nvidia to dedicate a full day of the conference to the topic.

Additionally, Nvidia is making strides in the PC market with its own central processor chips, a move that could challenge Intel’s remaining stronghold.

“It could eat into what’s left of the Intel market,” said tech analyst Maribel Lopez.

Huang’s keynote on Tuesday will set the tone for Nvidia’s next chapter as it navigates a rapidly evolving tech landscape.