Jensen Huang spent 33 years building Nvidia. His influence is easy to see. It’s harder to put a price on it. How much of what investors have been willing to pay for Nvidia stock came down to their confidence in him?

On November 12, 2026, Huang resigned as Nvidia’s CEO, president and a member of its board. Colette Kress, Nvidia’s CFO since 2013, will serve as interim CEO while the board considers internal and external candidates.

For me, the investment question is the same one I had before Huang left. If I buy Nvidia stock at this price, do I think it’ll outperform the other places I could put that money? I’m a bit skeptical. But I also think investors will keep paying a premium if Nvidia keeps delivering the earnings they expect. Huang’s resignation, on its own, does little to change my view.

The day before Huang resigned, Nvidia closed at $173.73, roughly 22 times the $7.91 it earned per share over the previous four quarters. With quarterly revenue up 106% year over year, I can see why someone would call that price reasonable. What gives me pause is that Nvidia can earn a profit selling chips for a data center long before its customer knows whether the investment paid off. At 22 times earnings, I’d want more confidence that those customers will come back for another order. 

The part I keep coming back to is what happens when it’s time to re-up. By the time a cloud provider considers a second data center, it has real numbers to work with. How often were the chips in use? What did customers pay to rent that computing power? After paying for the chips and running the facility, did the customer earn enough to build again?

A repeat order tells me more when I know why the customer placed it. If the first facility earned enough to make a second one worth building, I’d feel better about how much of Nvidia’s earnings I could count on over the next few years.

Even a customer that earned a good return on its first build still has to fund the next one. In August, Colette Kress said Nvidia expected roughly a quarter of the following year’s business to come from AI labs it planned to help finance. What I’d like to know is how much those labs would order, and how soon, if they had to secure that financing on their own. 

A customer could be happy with every Nvidia chip it bought and still push its next project back a year because borrowing got more expensive. Nvidia might eventually win the whole order, but the earnings investors expected from it would come later. That uncertainty was already there with Huang in charge, and it’s one reason I thought the stock was expensive before he left. 

Nvidia still makes some of the best AI chips a customer can buy. Huang’s departure raises questions about what Nvidia will build a few years from now, but the chips customers have already chosen can still do the same work today.

Take Amazon. In August, Nvidia said Amazon Web Services was deploying an additional two million GPUs through the second quarter of Nvidia’s fiscal 2029. If I’m Amazon, I’ve already decided those systems are a good fit. My first question after Huang’s exit is whether they’ll arrive on schedule. As long as Nvidia delivers the performance AWS planned around, I see little reason for Amazon to change course on that rollout.

Amazon’s next project is a different decision. If a competitor can run the same workload for less, Amazon has a reason to take a look. The question is how much it would actually save after its engineers adapt the software and test the new chips. Delivery matters, too. If a rival can get the project up and running six months sooner, that head start could be worth more than a discount on the hardware.

That’s how I could see Nvidia losing business over time. Amazon could finish its current rollout and still give a competitor the next order. For now, I’d expect customers to keep choosing Nvidia wherever its chips offer the best value for the work they need to do. The risk I see in losing Huang is that a few years from now, when customers compare their options again, Nvidia may have given its competitors time to catch up.

CoreWeave is where I take the argument about losing Huang more seriously. A cloud provider can decide Nvidia makes the best chips for its next data center and still have a hard time convincing a lender to finance it. One of the first questions I’d ask as a lender is who’s going to pay for all that computing capacity if the customers CoreWeave expects take longer to show up.

In September 2025, CoreWeave disclosed an agreement with Nvidia initially valued at $6.3 billion. If CoreWeave made the agreed computing capacity available and couldn’t sell all of it to customers, Nvidia was obligated to buy the remainder. In January 2026, Nvidia also invested $2 billion in CoreWeave shares.

How much of that depended on Huang personally? The deal terms tell us what Nvidia committed to. They leave open whether CoreWeave or its lenders would have had the same confidence in the project with someone else across the table. That’s where losing Huang could matter: a future deal might take longer to come together, or Nvidia’s next CEO might decide the potential chip sales aren’t worth putting that much money behind a customer.

Still, Kress has been Nvidia’s CFO since 2013. She knows what the company committed to and what it expected to get in return. I’d give that experience some weight before assuming future deals depended on Huang alone. What would change my view is seeing customers struggle to get their next projects funded even when they have buyers lined up and Nvidia is willing to back them. That would make me take a harder look at what Huang brought to those negotiations.

I’d feel differently about the price of Nvidia stock if its customers came back for a second build because the first one paid off. Say a cloud provider rents out enough capacity at a good enough margin that it can finance its next data center on the strength of those results. If it chooses Nvidia again, that gives me more confidence in what Nvidia can earn over several years. That’s the kind of evidence that would make me willing to pay more for the stock myself.

I could also be wrong about how much of Nvidia’s valuation depended on Huang. Suppose Nvidia keeps winning orders, delivers its next generation of chips on time, and earns roughly what investors expected. If, after a few product cycles, investors are still paying less for each dollar of Nvidia’s earnings while companies with similar growth prospects have held their valuations, I’d take that seriously. It would give me reason to think investors were paying for something I underestimated: their confidence in the bets Huang would make after the current roadmap ran out.

I thought Nvidia stock was expensive before Huang resigned, and I still do. If customers keep coming back because they earned a good return on the last build, I expect investors will keep paying more for Nvidia’s earnings than I would. Huang helped build that confidence. My bet is that it can outlast him, as long as customers keep seeing a reason to place the next order.