Jensen Huang Says AI Demand Is "Off the Charts." The Money to Pay for It Isn't

Chip stocks have been under pressure recently. During Thursday's trading session, the Philadelphia Semiconductor Index plunged by nearly 7%, and all 30 of the index's stocks finished the day in the red. Sector leader Nvidia (NASDAQ:NVDA) fell about 4.5%, as investors seem increasingly skeptical that the AI spending spree will last for years.

On that same day, Nvidia's CEO Jensen Huang said in an interview that demand from "neoclouds," companies that rent out Nvidia-powered computing capacity, is off the charts. But he also said that the money needed to build the capacity the market demands is far greater than what is actually available.

Image source: Getty Images.

Huang's proposed solution is to turn Nvidia's chips into what he calls a "financeable asset."

What exactly would this mean?

In simple terms, Huang wants to convince lenders to use the company's chips as collateral to secure loans for AI infrastructure, much like a vehicle secures an auto loan. Nvidia has been working with several investment banks for months to help mobilize more than $500 billion in financing for AI infrastructure.

Historically, chips have been a rapidly depreciating asset. Just think of how quickly the average laptop loses value and becomes obsolete, as a simplified example. But Huang argues that Nvidia's chips retain their value because they're productive, durable assets that can be transferred from one customer to another. Huang pointed out that rental prices for Nvidia's older H100 chips actually increased between Oct. 2025 and March 2026.

Some headlines have compared Huangis proposal to securing a mortgage, but as a financial planner, I think a car loan is a better comparison. A home is an appreciating asset. This is why home mortgage interest rates are generally lower than car loan rates. If the borrower can't pay back the loan, the lender can foreclose, sell the home, and (usually) recoup all of its money.

On the other hand, cars lose value over time, but lenders are more than willing to finance them because they know roughly how fast a car loses value. When it comes to AI chips, the question is whether their depreciation is predictable. Huang has said that Nvidia's chips can generate revenue for over a decade, but many in the industry are more skeptical; to be fair, computing hardware is typically depreciated over just three or four years. So, there are a lot of details to work out when it comes to the question of "what happens if something goes wrong and a loan isn't paid back?"

What this means for investors

Huang makes the point that demand isn't the problem and that helping its customers address their capital constraints could increase order volume.

On the other hand, this seems to make investors uneasy, and it's not hard to see why. Nvidia is effectively helping its customers borrow money to buy its products, and it has said it may guarantee part of the chips' future value to make that happen. The more of Nvidia's growth depends on its customers' ability to borrow money, the more exposed it will be if a credit crunch comes.

The bottom line is that Nvidia remains the dominant supplier, and demand isn't likely to disappear anytime soon. But like many investors, I'll be watching the credit markets closely. The soaring bond yields of the heavily indebted neoclouds tell us that the appetite for lending for AI infrastructure before there's real cash flow is declining. The AI build-out now depends heavily on borrowed money, and it looks like that won't change anytime soon. When you're deciding whether to buy shares of any AI infrastructure companies, keep this in mind as a risk factor.

Matt Frankel, CFP® has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Nvidia. The Motley Fool has a disclosure policy.