In his keynote address at the GTC 2024 AI conference, Jensen Huang named a few collaboration partners. Subsequently, the share price of these companies rose, yet Nvidia shares closed that day below the opening price.

On day two of the conference, when Huang and Chief Financial Officer Colette Kress answered financial analysts’ questions, investors apparently heard more of what they wanted to hear after the impressive yet abstruse technology-heavy keynote. Nvidia shares rebounded after taking a dip in early trading.

Huang estimated that companies operating data centers will spend around US$250 billion annually on accelerated computing components. He expressed confidence that Nvidia’s AI CPUs and software will grab a substantial share of this emerging market.

Nvidia is currently creating software for different industries and applications, lowering the threshold for adopting and using artificial intelligence. The company is also providing AI models and other software while charging its customers based on the computing power they use.

Huang expressed confidence in the company’s platform for AI services, saying “this is going to be a very large business”.

At the conference, Nvidia announced its new chip architecture “Blackwell,” which enables accelerated computing for two major AI tasks, namely training models and reasoning.

Huang told U.S. network CNBC that the Blackwell GPU is priced between US$30,000 and US$40,000, which is about the same as the previous generation Hopper GPU. Kress said that the new platform will likely become available later this year.

Following a series of statements by Huang and Kress, Nvidia’s share price, which had slipped 3.9 percent in early trading, rebounded to close 1 percent higher.

Its Rival AMD released last year its AI chip MI300 which at a price tag of US$20,000 is quite competitive cost-wise.

However, as far as technology and performance are concerned, Nvidia’s Blackwell GPU has further widened the lead over the competition. Still, when it comes to the data center business in the longer run, Nvidia will face the strongest competition from the tech giants. These large corporations strive to build their solutions as they are not willing to depend on outside suppliers like Nvidia.

Daniel Newman, CEO of the Futurum Group, points out that the rather lackluster performance of Nvidia’s stock price in the wake of the GTC 2024 AI conference probably shows that the hype about AI stocks has reached a temporary peak. “The market is looking far into the future. Everyone knows that Nvidia has grabbed an incredible market share, but as the new chip is born, investors are watching whether Nvidia will be able to maintain double-digit growth rates quarter after quarter,” says Newman.

As for Huang’s estimate that companies will spend US$250 billion per year to make their data centers fit for accelerated computing, Newman is more cautious: “The market wants to wait and see how big that market is going to be after all.”