The key to getting extra of the extremely coveted Nvidia graphics processing items important for advancing synthetic intelligence and machine studying is likely to be yellowtail sashimi.
At a gathering with analysts final week, billionaire Oracle cofounder and chief expertise officer Larry Ellison instructed the viewers that he and world’s-richest-man Elon Musk took Nvidia CEO Jensen Huang out to dinner at Nobu Palo Alto and “begged” Huang to present them extra GPUs.
“I’d describe the dinner as Oracle—me and Elon begging Jensen for GPUs,” Ellison recalled. “Please take our cash. Please take our cash. By the way in which, I obtained dinner. No, no, take extra of it. We’d like you to take extra of our cash please.”
The end result was a web constructive, mentioned Ellison. “It went okay. I imply, it labored.”
Ellison, whose famed fortune is estimated to be $206 billion, is a distinguished determine within the tech business and serves as government chairman of enterprise software program agency Oracle. He’s recognized for anticipating vital technological shifts, together with precisely positioning the $475 billion tech large to reap the advantages of the web in its early days. He has additional solidified the corporate by shifting it to capitalize on the surging want for cloud-based enterprise infrastructure. The corporate has maintained a robust relationship with Nvidia, which controls greater than 80% of the AI chip market, and is the prime alternative for firms engaged on AI.
Ultimately week’s assembly Ellison mentioned the aggressive panorama received’t hinder Oracle’s development. The only method of it, Ellison replied, is the phrase,“the race goes on.” He likened the present competitors in AI to a Components One driving race.
“You’ve obtained three folks on the rostrum, however there’s actually type of one winner,” mentioned Ellison. “Somebody goes to be higher at this than anybody else, and a number of persons are making an attempt and there’s a race.”
For its half, Oracle is making gargantuan investments in GPU expertise, significantly for AI functions. The tech large this month introduced its first quarter outcomes for fiscal 2025 and disclosed revenues have been up 7% to $13.3 billion, year-over-year. Income have been $2.9 billion. Oracle additionally disclosed that it has 162 cloud datacenters in operation or underneath building around the globe and the most important of its datacenters is 800 megawatts and can home “acres of Nvidia GPU clusters” used to coach large-scale AI fashions. Within the first quarter, Oracle signed 42 extra cloud GPU contracts, totaling $3 billion.
Equally, Tesla, the place Musk serves as CEO, depends on Nvidia GPUs to energy the supercomputers that prepare its neural networks for self-driving and assisted driving expertise.
In keeping with Ellison, being the primary to construct essentially the most succesful neural community on this planet “is an enormous deal.” Given the will to be on the forefront, many executives in AI are begging Huang for GPUs and constructing datacenters, he mentioned.
“Does anybody know what it prices to construct a frontier mannequin over the subsequent three years?” Ellison was referring to essentially the most superior, bleeding-edge AI methods.
“100 billion,” mentioned Ellison. “That’s going to get you within the sport.”
“Not lots of people and never lots of firms and never lots of international locations will take part.”
Nvidia and Tesla didn’t instantly reply to Fortune’s request for remark.
Information Sheet: Keep on high of the enterprise of tech with considerate evaluation on the business’s greatest names.
Join right here.