Larry Ellison and Elon Musk 'Begged' Nvidia's Jensen Huang for More GPUs for a Fancy Sushi Dinner
Yellowtail sashimi is the secret to getting more of the highly coveted Nvidia graphics processing units needed to advance artificial intelligence and machine learning.
In a meeting with analysts last week, billionaire Oracle co-founder and chief technology officer Larry Ellison told an audience that he and Elon Musk, the world's richest man, took Nvidia CEO Jensen Huang to dinner at Nobu Palo Alto and “begged” Huang to give them. More GPUs.
“I would describe the dinner as an oracle — me and Elon begging Jensen for GPUs,” Ellison recalled. “Please take our money. Please take our money. Anyway, finished dinner. No, no, take it more. We need more money please.”
The results were a net positive, Ellison said. “It turned out okay. I mean, it worked.”
Ellison, whose net worth is estimated at $206 billion, is a prominent figure in the technology industry and serves as executive chairman of enterprise software firm Oracle. He is known for anticipating significant technological change, with the $475 billion tech giant properly positioned to reap the benefits of the Internet in its early days. He further strengthened the company by capitalizing on the growing need for cloud-based enterprise infrastructure. The company maintains a strong relationship with Nvidia, which controls more than 80% of the AI chip market and is the main choice for companies working on AI.
At last week's meeting, Ellison said the competitive landscape won't hinder Oracle's growth. The easiest way to look at it, Ellison replied, is the phrase, “The race is on.” He compared the current competition in AI to the Formula One driving race.
“You've got three people on stage, but there's really only one winner,” Ellison said. “Somebody's going to be better at it than somebody else, and multiple people are trying and there's a race.”
For its part, Oracle is investing heavily in GPU technology, especially for AI applications. The tech giant announced its first-quarter results for fiscal 2025 this month and revealed that revenue rose 7% year over year to $13.3 billion. Profit was $2.9 billion. Oracle also revealed that it has or is under construction 162 cloud datacenters around the world, and that its largest datacenter will be 800 megawatts and house an “acre of Nvidia GPU clusters” used to train large-scale AI models. In the first quarter, Oracle signed another 42 cloud GPU contracts, totaling $3 billion.
Similarly, Tesla, where Musk serves as CEO, relies on Nvidia GPUs to power supercomputers that train its neural networks for self-driving and assisted driving technologies.
According to Ellison, being the first to build the world's most capable neural network “is a big deal.” Driven by a desire to stay ahead, many AI executives are begging Huang to build GPUs and datacenters, he said.
“Does anyone know how much it will cost to build a frontier model over the next three years?” Ellison was referring to the most advanced, bleeding-edge AI systems.
“A hundred billion,” Ellison said. “It will get you into the game.”
“Not many people, not many companies and not many countries will participate.”
An Nvidia spokesman declined to comment, and Tesla did not immediately respond fateits request.
Data sheet: Stay on top of the tech business with thoughtful analysis on the biggest names in the industry.
Sign up here.