The secret to getting your hands on more of Nvidia’s hugely popular graphics processing units, which are essential for advances in artificial intelligence and machine learning, may lie in yellowtail sashimi.
At a meeting with analysts last week, billionaire Oracle co-founder and CTO Larry Ellison told the audience that he and the world’s richest man, Elon Musk, took Nvidia CEO Jensen Huang out to dinner at Nobu Palo Alto and “begged” Huang for more GPUs.
“That dinner was like an Oracle dinner, me and Elon begging Jensen for the GPU,” Ellison recalled. “Take our money. Take our money. And by the way, I got dinner. Oh no, take more. Take more of our money.”
The results were overall positive, Ellison said: “It worked. I mean, it worked.”
Ellison, a tech industry luminary with an estimated net worth of $206 billion, is chairman of the board of enterprise software company Oracle. He is known for predicting key technological changes, including positioning the $475 billion tech giant to benefit from the early Internet. He further strengthened the company’s foothold by pivoting it to capitalize on the growing demand for cloud-based enterprise infrastructure. The company maintains a strong relationship with Nvidia, which controls over 80% of the AI chip market and is the first choice for companies working on AI.
At a conference last week, Ellison said the competitive environment won’t hinder Oracle’s growth. Ellison said the easiest way to frame it was, “The competition continues.” He likened the current race in AI to a Formula 1 driving race.
“There are three people on the podium, but there can only be one winner,” Ellison said. “There’s one person who is better than everyone else, and there’s multiple people trying and competing.”
Oracle, meanwhile, is investing heavily in GPU technology, especially for AI applications. The tech giant reported its fiscal first-quarter 2025 results this month, revealing revenue of $13.3 billion, up 7% year over year. Profit was $2.9 billion. Oracle also revealed that it has 162 cloud data centers in operation or under construction around the world, with the largest being 800 megawatts and housing “acres of Nvidia GPU clusters” used to train large-scale AI models. In the first quarter, Oracle signed an additional 42 cloud GPU deals, totaling $3 billion.
Similarly, Musk’s Tesla uses Nvidia’s GPUs to power supercomputers that train neural networks for self-driving and driver-assistance technologies.
Ellison said being the first to build the world’s most powerful neural network is a “big thing.” The desire to be at the forefront has many AI industry executives begging Huang for GPUs and building data centers, he said.
“Does anyone know how much it would cost to build a state-of-the-art model over the next three years?” Ellison was referring to cutting-edge AI systems.
“If you have $100 billion, you’re in the game,” Ellison said.
“There will not be many people, companies or countries participating.”
An Nvidia spokesman declined to comment, and Tesla did not immediately respond to Fortune’s inquiries.
Datasheet: Stay up to date on the latest happenings in the tech industry with in-depth analysis of the biggest companies in the industry.
Sign up here.
