Nvidia CEO Jensen Huang: Company Ready To Face AI Industry Shiftation

Nvidia CEO Jensen Huang, in a conference of miners (photo: x @nvida)

JAKARTA Nvidia CEO Jensen Huang confirmed that his company is in a strong position to deal with changes in the artificial intelligence industry (AI). This statement was made at the Nvidia annual developer conference in San Jose, California, on Tuesday 18 March.

In his speech, Huang explained that the AI industry is now starting to shift from the AI model training stage, where AI is trained with large amounts of data, to the inference stage, when the AI model is used to provide answers to users.

However, his statement was not enough to convince investors. Nvidia shares fell 3.4% after his presentation, while the semiconductor chip index (.SOX) fell 1.6%.

"Almost the whole world has miscalculated," Huang said on stage, wearing his signature black jacket and jeans. The number of computing needed by autonomous AI and reasoning-based processing is at least 100 times bigger than we thought last year.

Despite successfully dominating the AI market with a data center chip worth tens of thousands of dollars per unit, Nvidia faces new challenges. Startup AIs such as DeepSeek from China are reportedly able to develop competitive chatbots with fewer Nvidia chips.

Nvidia Introduces New Chip

Huang announced several new products, including the latest generation GPU chip, Blackwell Ultra, which will be released in the second half of 2025. This chip will have a larger memory capacity than the previous generation, so it can support larger AI models.

In addition, he also revealed a new chip system called Vera Rubin, which will replace Blackwell with a higher processing speed. The chip is slated for release in the second half of 2026. Nvidia also plans to launch its next generation of chips, Feynman, in 2028.

However, Blackwell's launch was delayed due to design defects that caused production problems. However, Huang said the demand for Blackwell remains "extraordinary."

"If AI takes too long to answer questions, customers won't come back. It's like a web search," Huang said. He also stressed that Nvidia chips are designed to provide intelligent and fast AI responses simultaneously.

In addition to the new chip, Huang also introduced an AI-powered private computer, DGX Workstation, which uses Blackwell chips. This device will be produced by Dell, Lenovo, and HP. With high specifications, DGX Workstation is a serious competitor to Apple's high-end Mac line.

"This is what a PC should look like," Huang said while holding the device's motherboard.

Nvidia also launched a new software called Dynamo, which is designed to accelerate the AI reasoning process and is available for free.

Collaboration With General Motors For Autonomous Cars

In addition to innovation in the AI and computing sector, Nvidia announced a partnership with General Motors (GM). The automotive company chose Nvidia to develop its autonomous car fleet.

With this series of innovations, Nvidia has further confirmed its dominance in the AI industry, despite the growing pressure from competitors and changes in the tech landscape.