Listen to the audio version of this article (generated by AI).
Hello, Reader.
“No one rings a bell at the top,” the old Wall Street saying goes. Often, things look strongest right before they turn.
Enter Nvidia Corp. (NVDA).
On Tuesday, Nvidia reported revenue of $115.8 billion, up 103% from a year ago, and adjusted earnings of $2.68 per share, versus expectations of $2.47. Data center revenue climbed 110% to $107.6 billion. Looking ahead, the company expects fourth-quarter revenue of approximately $130 billion, plus or minus 3%.
One particularly interesting detail: Nvidia’s fourth-quarter revenue forecast assumes zero data center computing revenue from China. So, its strong outlook isn’t dependent on getting Chinese sales back.
Nvidia’s results suggest that, at least for now, customers are still spending aggressively to secure computing capacity.
But while Nvidia’s earnings have dominated financial headlines, along with the scandal that led to Jensen Huang’s departure, many investors may be missing the actual story:
The era of Nvidia bankrolling the AI boom may be nearing its end.
This reminded me of a conversation I had with my father 40 years ago, when I was a brash and naïve 25-year-old. He and I sat down to discuss a prospective investment that required an initial commitment of $20,000.
“Just $20,000?” I replied dismissively. “That’s not a lot of money.” He shook his head and said, “It is if it’s your last $20,000.”
I’ve carried that line through a 30-year investing career, and it applies to companies as much as to people. Dozens of AI companies may have just received their “last $20,000” – and neither they nor their investors know it yet.
That’s why Nvidia’s financial backing may be the most important part of the AI story. If that support disappears, we’ll see which companies can stand on their own and which can’t.
So, today, I want to examine what happens when the company that helped finance the AI boom starts pulling back. Then, I’ll share how to identify the companies worth owning from those that may be too dependent on someone else’s balance sheet.
Nvidia: The Central Bank of AI
Nvidia has become much more than a chipmaker. It’s financing the AI boom, propping up its biggest customers, and helping keep the spending cycle alive.
The New York Times calls Nvidia the “Central Bank of AI,” and its latest filings help explain why. The company has:
- $108.5 billion in loan and lease guarantees, including $105 billion backing OpenAI’s Ohio data-center lease.
- $56 billion in commitments to rent cloud capacity and take on data-center leases for third parties.
- More than 10% of CoreWeave Inc. (CRWV) and guaranteed billions in its loans.
Together, those customer backstops total $164.5 billion – about 72% of Nvidia’s shareholder equity. Including future chip orders and other commitments, Nvidia’s total obligations could reach $530 billion, nearly three times the $184 billion reported just three months earlier, according to SemiAnalysis.
In other words, Nvidia is putting a huge portion of its financial strength behind the companies buying its chips and building AI infrastructure.
Morgan Stanley estimates Nvidia’s financial backing of the AI industry could reach $200 billion by the end of 2028, including $170 billion in guarantees. That’s a big commitment compared with the $56.6 billion Nvidia had in cash and investments at the end of July.
Ryan Cummings of Stanford’s Institute for Economic Policy Research estimates Nvidia’s financing arrangements would make it the country’s 10th-largest bank. He called Huang the driving force behind much of the AI buildout.
Under Huang, Nvidia’s influence extended well beyond the capital it supplied. The other half was confidence. Huang’s blessing could overcome almost any balance-sheet weakness. Thin revenue, heavy debt or reliance on one customer became less concerning once Nvidia backed a deal.
The former CEO acted as the AI industry’s godfather. If he said yes, lenders listened.
And whenever confidence in the AI story wobbled, he steadied it – with blowout sales figures, a rousing keynote, or a fresh check for a data-center project that looked shaky.
Nvidia’s backing also helped companies like OpenAI and CoreWeave borrow more cheaply. They could secure rates closer to what Microsoft Corp. (MSFT) or Alphabet Inc. (GOOGL) pays. Without that support, data-center startups have sometimes paid as much as 15% to borrow, compared with roughly 6% to 9% for Microsoft.
But here’s where the story gets interesting. Interim CEO and CFO Colette Kress never liked Huang’s open checkbook. According to The Wall Street Journal, she has long doubted the wisdom of having Nvidia finance AI startups and the industry’s expansion; and she now appears ready to slash those investments drastically.
Huang’s departure and Kress’ arrival come at a critical moment when other lenders are backing away.
Here’s why this matters…
Look Beyond the AI Borrowers
As I explained in last week’s Smart Money newsletter, banks are growing reluctant to finance more AI deals, even with guarantees from Alphabet and Anthropic PBC (ANT).
So, the AI economy needs more credit than ever, at the very moment it has lost the one backer who never seemed to say no.
This is where my father’s lesson hits home. CoreWeave, Nebius Group NV (NBIS) and other neoclouds have been spending their latest funding rounds as if more money will always follow. They may soon discover that their last funding round was their last easy one. They’ll find out when the next loan comes with a higher price – or doesn’t come at all.
If Nvidia pulls back, the entire AI financing machine could sputter.
But none of this means the AI buildout will stop. That’s why I’m sorting AI holdings into two groups:
Self-funders. These companies can afford to grow using the money they make from their own businesses. They may borrow, but they don’t have to. If lenders pull back, they can simply slow their expansion. Some may even pick up bargains when struggling competitors cancel projects or sell assets.
Dependents. These companies need outside money to keep growing. They rely on loans, financial guarantees or investors willing to keep buying their debt. Heavy debt, a handful of customers, and a business model that works only if the next financing round closes on time: those are the warning signs.
We want to invest in the first group. As the AI industry becomes less dependent on Nvidia’s financial backing, companies that can fund their own expansion may be better positioned.
This also points to a bigger opportunity: companies that don’t need to build the next generation of AI themselves. They just need to put it to work.
In my Life After Nvidia special event, I’ll look at which companies can survive if the easy money dries up – and which are positioned to profit from AI without spending billions to build it.
I am holding this event on Monday, November 23, at 10 a.m. Eastern. You can click here to reserve your spot now.
Nvidia can keep winning while the biggest profits from AI migrate elsewhere. I believe that the next phase of the AI boom may favor companies that don’t need to borrow billions to participate.
So, let the financial press dissect Nvidia’s earnings and debate who will succeed Jensen Huang. The question that matters for your portfolio is who will replace the credit line he extended to the entire AI ecosystem.
For now, the answer appears to be no one. And for dozens of AI companies, that means the last $20,000 may already be in the bank.
My advice: Don’t invest in companies that need someone else’s $20,000 to keep growing. Invest in the ones that know how to turn their own cash – and your capital – into something more.
Sign up for my Life After Nvidia event here.
Regards,
Eric Fry