At this point, there’s no denying that the problem with NVIDIA stock is sentiment, or rather distrust. Yes, that’s it. Distrust. There’s no other way to explain the stock’s lackluster reaction to NVIDIA’s quarterly report, not just after Wednesday night’s crash, but over the past few quarters. “Demand is growing parabolically,” CEO Jensen Huang concluded the call. He said demand was “soaring” last quarter. You almost feel sorry for him. He would soon run out of words to explain his request. Perhaps the most interesting part of Nvidia’s report is its new reporting framework. Specifically, we break down our data center business by hyperscalers (Amazon, Alphabet, Meta, Microsoft) and non-hyperscale customers. Listen, hyperscalers are important. They are spending hundreds of billions of dollars in capital investment to build new data centers. These were the main drivers of the past $200 worth of Nvidia stock gains. But what if it’s just an appetizer? On the conference call, Huang said Nvidia’s revenue will grow faster than hyperscale capex growth. So if hyperscale capex is what got us here, and Huang expects Nvidia’s revenue to outpace future hyperscale capex growth going forward, where will the upside come from? For everyone and everywhere else. We covered this in Wednesday night’s earnings analysis. However, let’s take a closer look at these non-hyperscalers. This group consists of dedicated AI computing providers called neoclouds (CoreWeave, Nebius, Iren, etc.). This also includes industrial companies and other businesses with on-premises computing infrastructure. Each country is building its own AI infrastructure, which Nvidia calls Sovereign AI. and other smaller AI players. Nvidia’s official name for this subsegment is AI Clouds, Industrial and Enterprise (ACIE). Juand said the ACIE cohort’s customers are “quite poorly understood,” in part because the market is fragmented. This could work in Nvidia’s favor in the long run, helping to address one of the biggest underlying concerns some investors have: hyperscalers developing their own custom AI chips. Huang said this ACIE opportunity is the consolidation of so many small AI players that there is no real demand for custom semiconductor solutions. He said these AI-native neo-clouds don’t want the complexity of designing a chip (which takes years and lots of money) or ensuring all parts of a data center work together properly. What they want is to run as fast as possible with the highest utilization possible. You need to be able to run any model and serve anyone, anywhere, anytime. To do that, Neocloud needs to be vertically integrated, and we need Nvidia to build vertically integrated data centers from hardware to networking to software. As Huang notes, Nvidia offers the most rentable architecture on the market with the highest total cost of ownership and ease of financing. “Obviously, we’re a huge part of that. We’re pretty unique in that we’re able to serve this industry. Our platform is built to be vertically integrated so that everything works. But then we break it down so people can build and buy in any configuration they want and put it back together the way they want.” It just keeps getting better and better. It’s not just that we have a large market share. That’s almost a 100% share, and a huge amount is inferential computing. That is, when the model is used after training. In other words, whenever you’re working with ChatGPT, it’s inference. Unlike training, which is more cyclical, inference scales with adoption and progresses upwards. And with a quick clip. This adds to the increased inference revenue the hyperscale subsegment is receiving from Anthropic, as the Claude model maker uses Nvidia silicon. “This segment is very fragmented and requires a fairly integrated, really well-integrated platform solution and a very large go-to-market. That segment, all guesses, 100% of it is Nvidia,” Huang added. This is very chewy. Let’s summarize. Huang believes NVIDIA’s ACIE customers could end up being dwarfed by hyperscalers, arguing that industrial and enterprise markets account for about $50 trillion to $80 trillion of the global economy (global gross domestic product was $111 trillion in 2024, according to the World Bank). Additionally, Huang said AI will help expand the size of that pie over time. So in this part of the market, Nvidia appears to be the only plug-and-play provider offering all the computing technology needed to stand up a data center (an AI factory, to use Nvidia’s terminology). Certainly, Nvidia may face some competition on the neocloud front. Soon, club name Alphabet and private equity giant Blackstone are developing an AI infrastructure company that runs on Google’s custom Tensor Processing Units (TPUs). Once this comes online, it will be a neocloud running non-Nvidia silicon. But when it comes to startups and other businesses that need help building their own high-speed computing infrastructure, Nvidia is the undisputed leader. NVDA 1Y Mountain Nvidia’s stock price performance over the past 12 months. Nvidia has growth drivers beyond its data centers. The company’s non-data center business accounts for less than 10% of its total revenue. This segment, now called edge computing, consists of gaming consoles, powerful desktops called workstations, personal computers, telecommunications, automotive, and robotics. Self-driving cars and robotics are included in the umbrella term of physical AI, and Huang is incredibly optimistic about the opportunities here. “Over the next five years, I expect the field of physical AI and robotics to grow at an incredible rate,” he said last night. What does this good news mean for the market on Thursday? The answer is somehow a 1.5% decline. This is head-scratching and reminds me of what we saw in March when NVIDIA announced a ton of good news at its GTC conference but its stock price didn’t move. A similar set of facts emerged here. Everything points to continued earnings growth, the stock is cheap, the story continues to improve, and the price movement does not reflect the fundamentals. Back in March, we said we needed to stay the course. That advice ultimately paid off in recent weeks leading up to the earnings, with the stock gaining momentum and hitting new all-time highs. I’m saying the same thing this time. We need to hold on to the stock and wait until the weight of earnings growth becomes too much for the naysayers to bear. Remarkably, NVIDIA has become a playbook for value stocks in the semiconductor industry, and one of the rules that keeps coming back to haunt us when violated is never to abandon value. To be clear, NVIDIA is not a value play, as its price-to-earnings ratio is slightly lower than its peers. No, it’s a great value play at this point. Please consider this. NVIDIA currently trades at about 23 times forward earnings, while its closest competitor, Advanced Micro Devices, trades at about 47 times, more than double that. In other words, Nvidia would need to rise about 50% and AMD would need to fall 25% to reach valuation parity. And Nvidia is a company that is attacking entire segments of the inference market unchecked thanks to its vertically integrated data center solutions. The bottom line? The market got it wrong on Thursday. Yes, this is not a company that has had a huge impact on its earnings in recent quarters, but instead tends to see its performance rise every time it releases a report. But that doesn’t mean it’s absurd to see a sell-off this quarter. The stock is trading at the lower end of its 10-year valuation range, even though the story is better than ever. Please be patient. If you don’t own the stock, now is the time to start taking a position because you’re essentially getting a quarter for free and the story is even better than we thought. That’s not likely to change any time soon. (Jim Cramer’s Charitable Trust is long NVDA. See here for a complete list of stocks.) As a subscriber to Jim Cramer’s CNBC Investment Club, you will receive trade alerts before Jim makes a trade. After Jim sends a trade alert, he waits 45 minutes before buying or selling stocks in his charitable trust’s portfolio. If Jim talks about a stock on CNBC TV, he will issue a trade alert and then wait 72 hours before executing the trade. The above investment club information is subject to our Terms of Use and Privacy Policy, along with our disclaimer. No fiduciary duties or obligations exist or arise from your receipt of information provided in connection with the Investment Club. No specific results or benefits are guaranteed.
