In this episode of Motley Fool Hidden Gems Investing, Motley Fool contributors Tyler Crowe, Matt Frankel, and Jon Quast discuss:

• Nvidia's earnings and guidance update.
• The winners and losers from Nvidia's earnings.
• CrowdStrike's earnings.
• Is CrowdStrike's stock a buy?
• Mailbag: Is local AI a hyperscaler problem?

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.

A full transcript is below.

This podcast was recorded on Aug. 27, 2026.

Tyler Crowe: Woah boy, did Nvidia's earnings deliver. 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 longtime Fool contributors Matt Frankel and Jon Quast. Guys, I think this has been the thing we've all been waiting for this week. It was Nvidia's earnings report. It came out after the close yesterday. Probably the most anticipated earnings report, even more so than the hyperscalers. I think it's fair to say that the report did not disappoint. Shares are up about 7% as we're taping this morning. Somehow, the company smashed earnings expectations and guidance expectations, which at this point, I'm starting to think Nvidia is arguably the most followed company on Wall Street. If the company can blast past expectations like this, I almost start to wonder, are Wall Street analysts not really good at their jobs, or was it just an incredible quarter here?

Jon Quast: Maybe both. Not all Wall Street analysts, but maybe some, but this is just an incredible quarter. It's every investor's birthday today with Nvidia. You look at the largest company in the world reporting 106% revenue growth at some scrappy start-up. It's utterly unbelievable, and then you look at the guidance for the coming year. It's guiding for 70% revenue growth on top of what it's reporting this year. We're talking tens of billions of dollars, and the thing with that guidance is that Wall Street was expecting good stuff. Wall Street is expecting 45% growth in the coming year, and VA is saying 70% growth. The thing with that is, it's indicating that it would even be higher if it could physically make the stuff, but saying that it is a supply-constrained outlook. In other words, if we could have more supply, then perhaps we could deliver more revenue growth in the coming year, we can't, so we're saying around 70%. Unbelievable, the top five hyperscalars spending roughly 800 billion in capex this year. Many investors believe that's been unsustainable that pace, but Nvidia saying they expect the hyperscalers to spend 1.3 trillion in the coming year, so 500 billion more than what they are going to spend this year. There's your headline numbers.

Matt Frankel: Companies that grow revenue at 106% year over year generally don't trade for 25 times forward earnings like Nvidia does. It's fair to say that the stock is pricing in an eventual deceleration. You mentioned the original estimate was for 45% revenue growth last year, which is a pretty big deceleration from 106%. Now that we're getting 70%, or they're saying 70%, which they have a good history of lowballing their estimates. That's the real reason the stock is rising. The company said itself that it can only satisfy about 70% of the demand it's seeing for its products. It's not the earnings themselves. Everyone expects Nvidia's earnings to be slightly better than the guidance and analyst expectations every quarter. In fact, Nvidia initially fell when the report came out before all those forward guidance numbers were revealed. There are margin compression concerns due to memory costs, but right now you're essentially trading about 300 basis points, or three percentage points, of your 75% gross margin in exchange for roughly doubling your revenue growth expectations going forward. The margins will be worth watching. Nvidia's margins have been expanding pretty much without fail for quarter after quarter, and that's no longer the case, but not much not to like about this report.

Tyler Crowe: It's strange. I'm a little bit at a loss for words for this, which makes for arguably the worst podcast media ever. But this guidance revision, this large, is obviously going to have a profound impact on the entire AI infrastructure ecosystem, not just Nvidia. We're seeing shares rise today and this idea that we're actually supply constrained versus, I don't know, maybe some of these new entrants like Google's, these custom-built chips that all the hyperscalers are building themselves. There was this concern that maybe demand might be there, but it is there. With that in mind, all of this supply coming to the market from Nvidia, from these other companies, and we're still supply constrained. We've covered the picks and shovels of the AI infrastructure build-out a couple of times here, but I want to challenge you guys a bit because this is the biggest story over the next couple of years. It's hard to avoid. What are some of the companies in this picks-and-shovels part of the industry that this revision signals to, that are going to be the winners and losers? Maybe let's focus on some of the ones that are a little lesser known here and really lean into our Hidden Gems theme here.

Jon Quast: Well, Tyler, I want to just point out what Matt said here regarding the gross margin concerns going forward for Nvidia, there's higher memory cost going into everything with AI, and that is going to impact some of Nvidia's margins. As you pointed out, maybe not a big deal with Nvidia, but let me just latch on to that higher memory cost thing. I think that that becomes a little bit problematic for a company such as HP. This is a loser here because of what's going on here. I'm not just saying that because HP stock is down today, but you look at consumer devices. Many of these memory companies, maybe not many, but some of them have pivoted away from the consumer market entirely because there's so much demand in the AI infrastructure trade. These memory companies pivoting exclusively to that.

Not only are memory prices going up generally, but now some of the companies supplying the consumer market have exited that. There's even a bigger shortage in the consumer part of the market. The memory costs are going to be a higher and higher percentage of the overall cost going into making a personal computer. The cost of a personal computer going up, that is really hard for a business like HP. You look at Dell, at least it has the tailwind from its AI server business that is supporting its growth. HP doesn't really have that. I would say it's a tough place to be if you're HP.

On the other side, let's talk about a hidden winner here. We're talking about potentially, I don't know how many extra GPUs that Nvidia is saying it's going to sell in the next 12 to 24 months. It's going to be a lot. Let's put it that way. GPU interconnectivity is a bigger and bigger trend. I look at a company such as Credo Technologies, ticker symbol CRDO. This is one of the several companies out there that can benefit from increased need in GPU interconnectivity. But when I look at Credo, one of the things that I do like about it is how small it is. Nvidia is saying 1.3 trillion potentially in spending in the next year from the top five hyperscalers. Credo Technologies only has 1.3 billion in revenue right now, and it's tripled in the last year, expecting 80% growth in the coming year. I believe this stock can be a good performer. Let's say if it reaches 5 billion in revenue within the next several years, that is just a drop of the gush of spending that Nvidia is projecting here. That is one that I think can be under the radar, maybe a hidden winner here.

Matt Frankel: I'll call out a few hidden winners and losers here, so a couple of winners. They're both roughly $15 billion market cap companies, and both are companies that most people haven't heard of. The first one's Onto Innovation, ONTO. They make the inspection tools used in making all these memory products that are fueling that supply challenge, and a between-the-lines takeaway is Nvidia's 70% expected growth rate next year. It implies that the company is going to need a lot more memory than expected, so that could be a positive tale in there. Another one's called IES Holdings, ticker symbol IESC. They design electric systems for data centers. We've talked about comfort systems. It's a similar story here, among other end markets they serve. If hyperscaler or capex goes from $800 billion to $1.3 trillion next year, power is a big constraint for data center development. Companies like this are going to have not only a lot to do, but a lot of pricing power. But on the losers' end,

Jon really hit the nail on the head when he said that any company that needs memory for any type of consumer electronics could be a loser. Although I love the company and its management team, I got to call out Garmin here, GRMN. They rely on third party memory for all their products and as much of a great job as they've done with companies like Nvidia scooping up more and more of the memory capacity, it could raise prices on Garmin devices or result in margin compression if they don't want to pass along price increases, it's likely to be a pretty temporary problem, but it's definitely one that's worth watching for companies like that.

Tyler Crowe: It'll be interesting to see if Nvidia is a chip company, but they outsource everything to Taiwan Semiconductor. I will be fascinated to see with the memory companies, as well as the chip manufacturers, the foundry companies, if we start talking about a significant expansion of production, because everywhere we look, we’re looking at these constrained bottlenecks, whether it be electricity, whether it be GPUs, CPUs, memory. Name your chip; we seem to be running short right now. I feel one of the stories we're going to be really following in the coming quarters is are we going to see significant expansion of capacity from these companies to actually meet this demand? Because otherwise, workers are just going to be stuck in this short-on-supplies thing, and a lot of these companies are going to have the winners and losers continue down this path unless there is some significant change. Coming up after the break, we got another earnings report coming up. We got CrowdStrike, who is also having an awesome day in the market.

Shares of CrowdStrike are up 18% as we're taping right now. After the company reported earnings, another great earnings report, similar to what we saw with Nvidia here. This pairs rather nicely with a lot of things that we've been talking about with AI because there was this story around AI and hacking and the idea of basically most of cybersecurity these days is going to be rendered obsolete by AI agents. But based on these earnings reports and Nvidia's guidance, it doesn't seem to be the actual issue here. It seems as though AI may be pushing customers to spend more instead of less. Guys, what did the report say?

Matt Frankel: Well, Nvidia's reaction was generally based on its forward guidance. CrowdStrike just reported a blowout quarter, period. Not because revenue grew 26% year over year, which was an acceleration. The real number is that net new growth in annual recurring revenue was up 51% year over year. That's a company record. That's something CrowdStrike has never been able to do, even in the much earlier phases of its growth ramping up. A net new AR that was about 17% higher than even management's own guidance. AI is clearly fueling demand for cybersecurity at a much stronger rate than experts thought. The remaining performance obligation, which is the technical word for the backlog, grew by 49% year over year to a little over $10 billion. Adjusted operating margin grew by 350 basis points. Management raised its full-year guidance significantly on both the top and bottom line. There were very high expectations going into this report. It's a highly valued stock, but CrowdStrike really delivered another great quarter.

Jon Quast: I love this not-so-subtle zinger from CrowdStrike CEO here from the conference call: AI is driving a clear divide between the cybersecurity companies that solve problems and those that compound problems. I'll let you decide which one CrowdStrike says it is itself, but what is so interesting here is, to Matt's point, this was a great new quarter for net new business during the quarter and the management team saying that this is as a result of the Mythos moment. If you remember Mythos coming out, this is one of those AI models out there supposedly broke containment and causing problems, and they had to pull it. Investors saw this initially as a threat to the cybersecurity space. CrowdStrike management saying this is actually going to be a benefit for us, and right now the numbers are backing that up. You see the timing of the Mythos being released and all that buzz, and then you see the uplift in CrowdStrike's business.

That said, I will put on my skeptical hat here because there are some recent changes in the last couple of years to CrowdStrike billing methods, and so all new customers are now coming in in this flex billing model. Whereas before, they signed contracts maybe for the year, and they picked which modules they wanted to adopt from CrowdStrike's roughly two dozen cybersecurity module options. Now you have a spending commitment. You can flex that into modules that you're experimenting with, flex away from the ones that aren't working for you. Then, within that time frame, you can re-flex. Otherwise, when you spend all the money that you had committed, you can just re-up at the same rates that you've negotiated with CrowdStrike. That really does allow for faster adoption. It allows for, rather than waiting for, let's say that you're a business and you see this Mythos moment, you're scared, you're wanting to try some things to protect yourself, whereas before you might need to wait til your contract was up or that whole thing.

Now you can just reflex, and you can experiment with things. Management pointing out that new customers are entering by the Flex business model, billing method. That's the one that they have to choose if they're a new customer. There's a significant uptake in the annualized recurring revenue right from the get-go. Maybe this is more of a billing thing that Mythos is showing us why this is such a good way to bill its customers, at least if you're CrowdStrike, but we'll see if that is actually great for both CrowdStrike and its customers long term.

Tyler Crowe: One of the more interesting things about AI, actually, is the way that it's changing the way a lot of software companies bill or charge how they're doing. We had this software as a subscription model, but now it's even starting to go almost on a usage basis, similar to what we're seeing with token usage with AI. Who knows, maybe one of the biggest innovations with AI is new billing processes here for these companies. CrowdStrike is another one of these darling companies. It's done incredible for investors. It's up 114% over the past year. There's plenty of Motley Fool members who've had recommendations on this around CrowdStrike who've done very well.

But I want to put the rubber to the road here for investors because another great quarter seems to continue. This is the crux of the challenge here is shares trade for 37 times sales. That means right now it would take 37 years for the company to generate its market cap in revenue. Now, granted that revenue stayed flat, the expectation is it would grow, but that's still huge. This was a discussion during the dotcom bubble with Sun Microsystems, where they're like, this isn't our earnings. This is our sales, before we have to actually deduct anything right now. This has been the story with CrowdStrike. It's a great business, it's growing a lot, but the valuation is crazy. With this valuation today, and everything you guys read from the earnings report and what you saw, is this stock an actual buy now, and if not, is there something within it doesn't have to be cybersecurity, but maybe the V-n diagram of what you could call CrowdStrike's tangential areas where investors should consider instead, if you think that this thing is just way too expensive?

Jon Quast: Well, Tyler, you're calling out valuation. I will say that I have sold stocks in the past based solely on valuation concerns. Every single time I have done that, I've lived to regret it in hindsight because, generally speaking, when the business is great, the market is great, everything is great, except for the valuation. Selling on that basis alone it's always been wrong for me. I tend to stay away from that more now. I really do try to sell businesses that are performing poorly or industries that I believe are under long-term threat. I'd let that be more of my deciding factor. If ever there was a company to not ignore the valuation entirely, but to push it way down your concerns list, I think CrowdStrike is one of them because it has demonstrated such an ability to compound its business over time, and I can't think of maybe a more important long-term industry than cybersecurity, one that's growing very nicely. CrowdStrike is one that, if your only concern is valuation, I'd hit pause on that because it really is a phenomenal business and industry.

I will say, though, if you are concerned about the valuation here but want to stay in the same general area, it's not a direct competitor, but I would just point out Rubrik, RBRK. This is a cybersecurity company, but not a direct competitor to CrowdStrike, it trades at only 14 times sales. That's about a third of the valuation, and yet it is growing faster than CrowdStrike. I think that is something to look at. It does generate free cash flow. It's got a pretty good balance sheet, so that would be one I'd look at.

Matt Frankel: I'm going to push back a little bit. If the pandemic era, SPAC bubble, and the Dublin high-growth tech stocks taught me anything, it's valuation always matters, at least a little bit. Contrary to what Jon just said, there are several stocks that I regret not selling on ridiculous valuations during that era. Just to put the numbers in perspective, CrowdStrike they've previously targeted 20% long-term revenue growth in perpetuity. At the current valuation, that tells us that the stock trades for about seven times the revenue it will generate in a decade from now. On one hand, growth is accelerating, which, you're right, is rare at this scale. The forward book of business is growing faster than current revenue. That implies acceleration in the near term.

That AI cybersecurity thesis is something that I completely believe in, but the math here is brutal. Even if, let's say, CrowdStrike's revenue were to triple over the next five years, which I think would be a stretch, the company would still be at about 12 times sales at the current price. There's a lot not to like here, and I know Tyler is going to be with me on this one. Their stock-based comp is massive enough to completely wipe out that operating income that we talked about. The company actually reports an operating loss on a GAAP basis. Their diluted share count grew by almost 5% over the past year. That's even after some buybacks to help offset it, and I would expect a similar rate going forward. Even though they're a leader, they're not without competition.

To modify a Warren Buffett quote that I love, CrowdStrike is a wonderful business at an uncomfortable price right now. I wouldn't start a new position here today. If you absolutely want to, it's definitely a great case for dollar-cost averaging into a position. But as an alternative, I recently bought Zscaler. They partner with CrowdStrike, a lot of the same tail ones, but a much more palatable valuation right now.

Tyler Crowe: The stock-based compensation, and this is not just a Crowd trick thing. It is something that I have always struggled with when it comes to evaluating businesses is when there is that much stock-based compensation that wipes out earnings or operating profits in that way, it's hard to assess the success of this business over the long term because so much bid is going to the employees rather than the shareholders. It will be a story worth following. With a valuation this high, I know that investors are going to be thinking about that over the long term. Coming up for the break, we're going to dip into the mailbag specifically related to AI.

ADVERTISEMENT: At McDonald's, feeding the little ones just got easier with a Happy Meal for just 299 on the app. That's long journey stop-offs and move-in fees sorted. Just a little spin for some big smiles. Some fun, some food. It's all inside this Happy Meal. Until 1st September from 11:00 A.M. Only on the app, not on delivery. One use per day, a maximum of four Happy Meals per redemption. Price may vary. Treacherous ring roads lie ahead on our mission to reach nanos. There is Mackey's for Black Pepper and Garlic McNuggets. Navigation, rerouting. The main mission can wait. Take a side mission to McDonald's. Available until 29th September from 11:00 A.M. Whilst stocks last, subject to availability. Participation varies.

Tyler Crowe: Everyone, quick reminder, if you want to get a question into us and have it read on air, email us at [email protected]. That's podcast with an S. I've also put the email in the show description, so you can get it there as well. Remember, keep it short, keep it Foolish, and we cannot give personal advice. Today's question comes from John from New Jersey, and the question is, we have had multiple discussions about whether the money invested in data centers will be worth it, and that's a very fair assessment of what we've done here. One of the threats to data centers is the idea that many of the new models can be run locally, not necessarily like cloud and OpenAI, but a lot of the Chinese models that we're seeing these days. One big advantage of these models is that all of your company's data will remain local and will not be sent to another company to possibly be sold on. The newly announced Apple computers, personal computers. I think it is the new Mac Mini, are actually designed to run AI locally. This gets to Jon's question, how much of a threat is local use of AI to the actual AI cloud companies?

Jon Quast: I can't tell you how much I love this question. I feel I have been a local AI crusader for months now, and I've been all by myself. Now John from New Jersey is standing with me. Johns got to stick together, so I'm really thankful to talk local AI. Just briefly, Apple, many products in particular, have already been used for local AI, and Apple leaning into that with its next generation, but let me answer this question directly. I don't think that local AI is a threat to the cloud AI industry. I think that there are different use cases. I think that more and more people, individuals, are going to want to run some local AI. Running a model that is personalized to me on my stuff in my home, I think that's going to be a growing trend. There's a lot to unpack with that. There's still a lot of learning how to structure it.

A lot of people out there are trying to solve for this, but beyond individuals, I think that hospitals, schools, government agencies, they're all going to want to run more and more local AI for obvious reasons. They're not going to want all that stuff in the cloud. Yet I do think that there's still a use case for cloud AI in the future, because that's where a lot of the innovation is going to be happening. A lot of running your own stuff on your own hardware, there are challenges to that. I think that SpaceX's new Grok bot really is making it more usable, perhaps driving an adoption curve. Definitely watch that, but I still think there's going to be innovation happening and things that you're going to need a cloud AI for. I think it's going to be both in the future. I think that we already have so much money being spent going into the data centers, going into those. But I think there's going to be a layer on top of that, which is your own local AI at home, or in your place of work, place of business, your school, whatever. I think it's going to be both in the future, so I don't see this as either/or.

Matt Frankel: I generally agree with Jon. It's really a case of a rising tide lifts all ships here when it comes to AI. Data centers and on-device AI are two totally different things to solve, two totally different problems. Plus, data centers have a few key advantages that are going to help them continue to grow. We've already mentioned the memory problems several times in this episode. The AI boom has made memory for local AI much more expensive, just to name one thing. Data centers, their operators have a lot more room to win the memory price awards. They have more money. They pay more if you're investing hundreds of billions of dollars. You get priority. Local inference it's bound by memory capacity, not by raw compute power. That's not the constraint here. That is a big challenge, but while local AI is not much of a threat to cloud-based AI, to tie it into our earlier discussion, I do think it could be one of the best things that ever happened to cybersecurity. Think about it, if all of your employees' local devices are running their own AI models, there will be thousands of new AI endpoints that you need to secure inside enterprise networks. Definitely could help companies like CrowdStrike and Zscaler going forward, just to tie it into what we were saying earlier.

Tyler Crowe: There's a real irony here of the entire 2010s was the whole movement of migration to the cloud, and every single business wanted to migrate to the cloud. Now we're talking about everything going back onto local servers, local hardware, and keeping your own stuff in-house. Everything seems to come full circle. Well, guys, that's all the time we have for today.

As always, people in the program may have stocks they talk about, and The Motley Fool may have formal recommendations for or against, so don't buy or sell stocks based solely on what you hear. All personal finance content follows Motley Fool editorial standards, and it's not approved by advertisers. Advertisements or sponsored content are provided for informational purposes only. See our full advertising disclosure, please check out our show notes. Thanks for producer Dan Boyd and the rest of The Motley Fool team. For Jon, Matt, and myself, thanks for listening, and we'll chat again soon.