Larry Ellison is the chairman of Oracle (NYSE: ORCL), which is at present constructing among the quickest and most price environment friendly information facilities on this planet for growing synthetic intelligence (AI). Elon Musk, alternatively, runs Tesla (NASDAQ: TSLA), which is constructing AI-powered self-driving software program for its electrical autos. He additionally runs SpaceX, X (previously Twitter), and a brand new AI start-up known as xAI.
Ellison and Musk want tens of 1000’s of graphics processors (GPUs) for his or her information facilities with the intention to carry AI to life, and Nvidia (NASDAQ: NVDA) provides the most effective chips within the business.
At Oracle’s monetary analyst assembly on Sept. 12, Ellison advised the viewers that he and Musk just lately went to dinner with Nvidia CEO Jensen Huang on the Nobu restaurant in Palo Alto. The 2, who’re among the many richest folks on Earth, discovered themselves begging Huang for one thing cash merely cannot purchase in the meanwhile. Here is the way it went down.
The arms race for GPUs
Oracle at present has 162 information facilities both stay or below development, however it believes that quantity may ultimately prime 2,000 as a result of the demand for computing energy from AI builders is hovering. A few of Oracle’s largest information facilities characteristic clusters of greater than 32,000 GPUs, however subsequent 12 months the corporate will provide a cluster of 131,072 GPUs from Nvidia’s newest Blackwell lineup.
Oracle designed distinctive RDMA (random direct reminiscence entry) networking expertise that may transfer information from one level to a different extra rapidly than conventional Ethernet networks, and since builders pay for computing energy by the minute, this will considerably scale back prices. That is why main AI start-ups like OpenAI, Cohere, and even Musk’s xAI are utilizing Oracle’s infrastructure.
In its current fiscal 2025 first quarter (ended July 31), the Oracle Cloud Infrastructure (OCI) phase generated $2.2 billion in income, a whopping 45% bounce from the year-ago interval. Nevertheless, it could possibly be rising even sooner if not for provide constraints — in different phrases, Oracle merely cannot get its palms on sufficient GPUs for its information facilities.
Not solely is Oracle battling different cloud giants like Microsoft, Amazon, and Alphabet for GPU allocations from Nvidia, however tech corporations like Tesla and Meta Platforms are additionally absorbing provide to develop AI for their very own functions. Tesla is making an attempt to carry a cluster of fifty,000 GPUs on-line this 12 months to boost its self-driving software program, which requires a considerable quantity of computing energy.
Meta, alternatively, used round 16,000 of Nvidia’s flagship H100 GPUs to coach its Llama 3.1 giant language mannequin (LLM), however the firm plans to extend its capability to a mind-boggling 600,000 H100 equivalents by the top of this 12 months. That may pave the way in which for Llama 4, which CEO Mark Zuckerberg says may set the benchmark for the business in 2025.
Ellison and Musk are begging for extra GPUs
Please take our cash … take extra of it. You are not taking sufficient. … We’d like you to take extra of our cash. Please.
— Ellison’s and Musk’s feedback to Jensen Huang over dinner, in response to Ellison.
Ellison and Musk have been virtually begging Huang for extra GPUs, however no amount of cash on this planet should buy the numbers they require proper now as a result of Nvidia merely cannot sustain with demand. Oracle and Tesla aren’t even Nvidia’s largest clients!
Oracle spent $6.9 billion on capital expenditures (capex) throughout fiscal 2024 (ended April 30), and it expects to spend double that in fiscal 2025. Many of the cash will go towards shopping for chips and constructing information facilities. Tesla plans to spend over $10 billion on capex this calendar 12 months on the entire, which may even go towards the 50,000 GPU cluster I discussed earlier.
These numbers are modest in comparison with what different tech giants are spending. Microsoft allotted $55.7 billion to capex throughout its fiscal 2024 (ended June 30), and it plans to spend much more in fiscal 2025. Amazon’s capex spending, alternatively, may prime $60 billion in calendar 2024.
Due to this fact, it is no shock that Nvidia generated $26.3 billion in information middle income throughout its current fiscal 2025 second quarter (ended July 28), a 154% enhance from the year-ago interval. Ellison says the wave of AI spending may proceed for the following 10 years as corporations and nation states battle for tech supremacy in terms of AI, so Nvidia’s information middle income most likely has loads of development left within the tank.
Must you make investments $1,000 in Nvidia proper now?
Before you purchase inventory in Nvidia, take into account this:
The Motley Idiot Inventory Advisor analyst crew simply recognized what they imagine are the 10 finest shares for traders to purchase now… and Nvidia wasn’t one among them. The ten shares that made the minimize may produce monster returns within the coming years.
Take into account when Nvidia made this record on April 15, 2005… for those who invested $1,000 on the time of our advice, you’d have $715,640!*
Inventory Advisor gives traders with an easy-to-follow blueprint for achievement, together with steerage on constructing a portfolio, common updates from analysts, and two new inventory picks every month. The Inventory Advisor service has greater than quadrupled the return of S&P 500 since 2002*.
*Inventory Advisor returns as of September 16, 2024
John Mackey, former CEO of Entire Meals Market, an Amazon subsidiary, is a member of The Motley Idiot’s board of administrators. Suzanne Frey, an government at Alphabet, is a member of The Motley Idiot’s board of administrators. Randi Zuckerberg, a former director of market improvement and spokeswoman for Fb and sister to Meta Platforms CEO Mark Zuckerberg, is a member of The Motley Idiot’s board of administrators. Anthony Di Pizio has no place in any of the shares talked about. The Motley Idiot has positions in and recommends Alphabet, Amazon, Meta Platforms, Microsoft, Nvidia, Oracle, and Tesla. The Motley Idiot recommends the next choices: lengthy January 2026 $395 calls on Microsoft and quick January 2026 $405 calls on Microsoft. The Motley Idiot has a disclosure coverage.
You Will not Consider What Larry Ellison and Elon Musk Mentioned to Nvidia CEO Jensen Huang was initially revealed by The Motley Idiot
