When Larry, Elon and Jensen Met, This Happened

In Palo Alto recently at Nobu restaurant, Larry Ellison, CEO of Oracle, met with Jensen Huang, CEO of NVIDIA and Elon Musk.

If you could have been a fly on the wall listening in on the meeting, you might imagine that you could glean some valuable insights as to where the world is headed, what the future of artificial intelligence will look like, and who will benefit most.

As it turns out, we got the next best thing, a discussion with Larry Ellison about how, in fact, that meeting went, and it is captured here:

What did @larryellison say about AI?

“I went to dinner with @elonmusk & Jensen, and would describe the dinner as…me & Elon begging Jensen for GPUs.

Please take our money…No, no, take more of it. You’re not taking out enough. We need you to take more of our money, please”

— Financhill (@financhill) September 14, 2024

When Larry, Elon and Jensen Met

The meeting of the heads of Oracle, NVIDIA and X led to one clear outcome. A fight among the top companies in the world is taking place now to get ahold of as many GPUs as NVIDIA can provide.

Larry went on to describe AI as being a crucial race to win for the top companies in the world. Much like a Formula One race, he emphasized how important it will be to come first. Failing that, being best at something is going to be crucial to long-term survival.

In order to compete, Larry stated, the cost is astronomical. He estimates the minimum cost to enter the AI race is $100 billion when you factor in all the GPUs across literally acres of land demanding massive compute power. It’s so intense that nuclear power reactors are needed to generate sufficient energy to meet the needs.

But clear as day in his message, too, was the fact that there will be no easing up on demand for GPUs anytime soon among any of the top firms, whether that’s Microsoft, Alphabet, or Oracle. All of them have their hands out to Jensen begging for more compute power, meaning more GPUs, and that all boils down to one thing.

NVIDIA acts like a toll road to whom major technology companies must pay if they want to seriously compete in artificial intelligence. Absent NVIDIA’s GPUs, they simply cannot gain entry to the competition, and so when you hear about the massive spend rates on AI among the Magnificent 7, you can interpret those figures to mean that a substantial portion will make their way to Jensen’s firm.

So what does it all mean from an investor perspective?

Buy NVIDIA On The Dips

What makes investing hard is that most analysis looks to the past and extrapolates to the future. Just a few years ago, when NVIDIA chips largely served the gaming community, it would have been hard to imagine a day when a substantial portion of the capital expenditures of the Magnificent 7 would be allocated to buying its chips. Yet that’s precisely what has happened.

Revenues have soared from $4.6 billion in 2015 to $60.9 billion this year. Earnings before interest and taxes have skyrocketed from $759 million to $32.9 billion over the past year. Gross margins have soared from 55.5% ten years ago to 72.7% more recently.

With no signs of spending among the best capitalized technology firms in the world slowing, betting against NVIDIA now seems like a fool’s errand.

As NVIDIA’s profit and loss statement looks ever more impressive, so too does its balance sheet. From ballpark $4.5 billion in liquid reserves ten years ago, NVIDIA has closer to $26 billion now, and that in turn provides a lot of optionality. 

From an investor’s perspective, this can be boiled down to a simple idea: buy NVIDIA stock on the dips because the odds are high between now and 2029, in Ellison’s view, NVIDIA will continue to benefit the most.

What Companies Benefit The Most From AI Spending?

In the short-to-medium term NVIDIA should benefit the most from AI spending as Microsoft, Alphabet, X and Oracle invest heavily in GPUs, but over the longer term the Magnificent 7 will benefit most.

The reason the major technology titans are likely to benefit most over the long-term is that the applications they build today will have benefits that persist for the foreseeable future.

In the short-term, NVIDIA is essentially selling the picks and shovels to let them mine for gold. Over the long haul, the software built is likely to have persistent demand.

Ellison discusses the range of applications that will be disrupted are as wide as one can imagine now, whether it’s biopharma companies researching cancer cures or media companies creating AI movies. Virtually every aspect of our lives will be touched by artificial intelligence and yet only a very few companies, let alone countries according to Larry, will be able to afford to enter the game in the first place. 

So, if you’re looking to place a bet on AI over the next 5 years, the best approach may well be the one Buffett took with the airlines a few years ago. Rather than make a forecast that a single airline would win, he bet on the top four airlines and took a stake in all of them.

Similarly, a potential investment angle is not simply to view NVIDIA as the winner selling the picks and shovels today, but also to take a stake in Microsoft, Alphabet, and Oracle.

Apple is a top contender too, but its capex spend is substantially lower than its rivals, and arguably Apple knows that, regardless of which AI applications win, all of them will sit on top of its iOS platform, so it can afford to sit back and wait for a more obvious play to emerge before committing gigantic sums of capital.

---

The author has no position in any of the stocks mentioned. Financhill has a disclosure policy. This post may contain affiliate links or links from our sponsors.