How Nvidia Could Make $150 Billion Look Small.

In this episode of Motley Fool Hidden Gems Investing, Motley Fool contributors Tyler Crowe, Lou Whiteman, and Rachel Warren discuss:

• Nvidia’s $150 billion buyback plan.
• The case for a $500 billion buyback.
• Nvidia’s AI safety plan and what the market doesn’t like.
• What to make of all these new defense companies.
• Mailbag: Thoughts on Klarna?

To catch full episodes of all The Motley Fool’s free podcasts, check out our podcast center. When you’re ready to invest, check out this top 10 list of stocks to buy.

Transcript is below.

This podcast was recorded on Sept. 28, 2026.

Tyler Crowe: One hundred fifty billion, not that big. Motley Fool Hidden Gems Investing starts now. Welcome to Motley Fool Hidden Gems Investing. I’m your host, Tyler Crowe. Today, I’m joined by long-time Fool contributors, Rachel Warren and Lou Whiteman. We are going to get into the promise and the perils of investing in the defense and security sector. We’re also going to hit the mailbag. But we want to start the day with, I guess you could say, the biggest news story, and that was Nvidia. They had a couple of press releases today, one on AI safety and for the investors, perhaps the bigger one is the board of directors announced a $150 billion share repurchase agreement. As a press release says, the biggest in history and gives the board about $250 billion to buy back shares right now. Now, guys, I’m going to give a hot take, and I want to get your quick reactions to this. I think it’s way too small.

Lou Whiteman: I mean, look, they’re finally off the 1-cent-a-share dividend, which I didn’t understand that either. But the law of large numbers just even that’s thrown out of proportion here, but nothing is big enough. One hundred fifty billion? Yawn.

Rachel Warren: I mean, it’s the biggest in history, but it’s also only around 2% of its market cap, so you have to right-size that for the company we’re talking about. The other way to look at it is Nvidia is generating tremendous amounts of cash flow from the AI infrastructure buildout. They have more capital than they can logistically reinvest back into R&D. I think that this is a way we’re seeing that play out.

Tyler Crowe: I would make the argument that the announcement should be at least triple this. I know that sounds like crazy and probably put me on CNBC.

Lou Whiteman: Well, that’s crazy.

Rachel Warren: I think that’s definitely part of it. I don’t think the timing is a mistake by any means. I think that was probably very well coordinated up above, but this is a real industry issue. We’re seeing advanced A models, as you noted, from a wide range of labs, as well as the big hyperscalers that are repeatedly breaching security. Sandboxes, obviously, there’s the high profile Hugging Face incident. But we’re also at a time where CEO Jensen Huang has been very clear on the approach he wants to take here. He doesn’t want to wait for government regulation. He’s actually back against that. He’s really framed AI safety strictly as an engineering problem that can be managed at the hardware and platform layer.

ust to talk a little bit about what they’re doing, Nvidia is introducing an open source framework, it’s called OpenShell. They’re also introduced a hardware watchdog called Sentry, essentially the idea is these will help quarantine rogue software within milliseconds. OpenShell is interesting. It’s open source, as its name indicates, it’s designed to support rival platforms from ARM and Intel. You can see how Nvidia is trying to establish a universal industry standard for safety. They want to lower the risk barrier for enterprise adoption before maybe you see more corporate security teams or regulators freeze the deployment of autonomous AI agents. I think that they’re very much trying to get ahead of the curve here.

Lou Whiteman: Back to the investing question. I think they’re couching bad news with good. But also, look, this is part of a trend, Nvidia has run out of ways to wow the market. Jensen increasingly he looks like the greatest showman. He knows how many times he now goes on stage to just make bold pronouncements. The stock has done nothing basically, since the spring. Amazing earning reports are just being met with yeah. I think, look, this is a great business, but the perception is, whether it’s AI doomerism they’re trying to push back against or just investor boredom, Nvidia has run out of ways to wow or excite investors. I think the big buyback, this is just another maybe failed attempt to say, remember us, we’re still really good.

Tyler Crowe: Well, guess what? If you’re the CFO and you really want to wow investors, $500 billion buyback, that’s going to be by pitch to them. But to that point though, Nvidia trades for about 18 times its earnings estimates for next year. It’s the cheapest in a decade. A nice or a little quib I like the market is irrational, but it isn’t dumb. There is some risk building here that maybe we’re not seeing. What may those risks be that investors might want to consider that are maybe hidden that most people aren’t thinking, of course, there’s cyclicality in this industry, but what are some of maybe people aren’t considering this one as much?

Rachel Warren: I think there’s a few things to consider. I mean, one is the circular funding. This isn’t hidden, this is something that’s been talked about a lot. I mean, we look at that number that Invidia just announced today. They’ve disclosed that Anthropic’s aggregate contracted value across NVIDIA-backed cloud entities has scaled above $180 billion alone. That’s just in their arrangements with Anthropic. Nvidia’s balance sheet includes 13 public stakes, over 200 private stakes, $20 billion in data center lease commitments. They’re operating essentially like a central bank for a lot of the AI industry. Part of that is a function of necessity. Obviously, the idea is that that flows back into the growth story for Nvidia.

But there’s that circular funding mechanism that I think is concerning depending on if or when the build-out slows down. One other thing to note, hyperscalers are writing off these AI GPUs on aggressive 3-5-year life cycles. They’re operating under the assumption that there’s going to be these constant multi-billion dollar upgrades. But what if you have a GPU that’s a few years old that ends up being perfectly adequate for 80% or more of enterprise inference workloads? That upgrade cycle could be much slower. If that happens, you could see companies like Nvidia face a bit of a cliff or at least a slowdown in growth, even if it’s a distant one. I mean, essentially, this idea, not that the technology fails. If the existing hardware works too well, you could start to see Nvidia frozen a little bit out of its own upgrade cycle. That’s the extreme case, but I do think it’s something that if you’re an investor in Nvidia, if you’re looking at Nvidia right now, these are a couple key factors you need to be incorporating into your investment thesis.

Lou Whiteman: I think there’s a pretty not-extreme straightforward case to say that the best is now behind them for the AI cycle. Not that it can’t be really good for a long time. But Google, Amazon, Microsoft, Meta, OpenAI, they’re all at least trying to make their own chips. No one is going to stop doing business with Nvidia, but Wall Street pays for growth, if there’s more competition coming in, I didn’t even mention the other chip companies, that is going to put a cap maybe on upside growth. Also, if the frontier models don’t matter as much, and I think evidence is growing that they don’t matter as much, then does having the latest, greatest, newest, most expensive chip matter as much at the scale we hoped? You say it’s boring, but it does matter. Cyclicality is a thing. As recently as 2023, Nvidia was generating earnings per share under a dollar per year; now it’s at eight. It may look cheap on a price-to-earnings basis, but there is a question of is this level of earnings per share, how long could it go on? If not, that multiple looks a lot different.

Tyler Crowe: Well, you know what brings up that earnings per share? Five hundred billion dollars by ABAC. I am going to solve this one into existence. Coming up for the break, we’re going to talk about the defense and security industry.

REDLattice is a cybersecurity company that’s going public via SPAC. It expects to merge with a company that’s called Bold Eagle Acquisition Corp., and it’s going to conclude sometime around the end of 2026. It’s expected to go public at around 1.25 billion. I would say the layman’s way to describe it is it’s a competitor to Axon Enterprise and Cellebrite in cracking encrypted devices and other sensitive information processing. There’s quite a few of these companies starting to pop up in the defense industry and the company itself isn’t why I bring this topic up, though, it’s making its rounds on the financial media rings. I think that this is indicative of a shift in the market’s attitude towards defense tech and somewhat related space companies. For much of the 2000s, 2010s, the defense industry was consolidating. It was fewer players and those giants to me, they weren’t that far removed from utility companies in terms of growth, earnings, what have you in terms of the returns.

For a myriad of reasons, that’s changed, and there is bound, I think, to be a fair share of companies going public in this space, some of which are going to be lower quality companies, very similar to what we saw with SPACs in 2021 in the fintech technology consumer tech region. Lou, you’ve done defense for a little while. Rachel, you’ve given some thoughts on this as well. When you’re looking at IPOs for defense-related companies, we’re going to see a lot of them. What are some of the specific traits to defense investors should look for?

Lou Whiteman: Well, let’s start real quick outside of defense. Just why do companies go public via SPAC? The big thing is, is that a SPAC is technically a merger, and SEC rules give you a lot more leeway to talk about the future in a merger than they do in an IPO. That matters if you don’t have much of a past to talk about. We are talking about early-stage companies that really want to talk about the future. That fits in real well. As far as defense-specific, the Pentagon has been talking about diversity a lot, but it doesn’t want to deal with these companies; it really doesn’t. What am I looking for? I’m looking for on RAMS. I’m looking for companies that work with companies that have direct access, whether it’s the primes, whether it’s this new generation. But it is not easy to get in the cool guys’ club at the Pentagon. A lot of these, even if their technology is better, are just not going to generate the revenue that they hope to if they aren’t in the room.

Tyler Crowe: Lou, I’m going to stop you for a quick sec. Just for everyone who’s not a super defense, sicko, what is a prime? Because I am sure there’s some people thinking, what is that?

Lou Whiteman: The prime contractors are, well, I mean, for the most part, we think of them as Lockheed Martin, Northrop Grumman and the big mass of guys, but they’re smaller ones, too, but basically, they are the ones in the room. They are the ones with direct contacts and contracts with the Pentagon. They’re called primes because literally, the Pentagon does business with them, and then they just dole out revenue to other companies that work with them. Another thing you’ve got to consider here is most of these defense tech companies just have some novel tech and it looks like wow, and it looks like they’re the only ones who can do this. There is a ton of government-funded R&D going on with the inside the established contractors. Because it’s classified, they don’t talk about it much and they don’t have a need to talk about it the way you do if you have an IPO. I’ve seen numerous companies basically fail because what looked to be unique or looked to be something they could only do was actually yesterday’s news on the inside.

It’s really, you can’t do due diligence here because so much of it is classified, but just be aware that if the first thing you think is, why did anyone think about this? That might be a yellow flag. Also, and again, you just can’t say this enough that just the idea of how you get in the room. For most of these companies, working with the government is a core competency. It’s really hard to deal with government bureaucracy and most companies can’t. It is a skill to be able to navigate this. I don’t want some Silicon Valley dudes with an idea. I want at least someone in leadership, whether it’s a chairman or something that has government related experience. Related, you need a big employee base with security clearances. If you don’t, you can demo whatever you want, but you’re never going to get to interact with the defense chain. Those are just some yellow flags to look at when you’re looking at these.

See More

Source link