Nvidia shares fell on Wednesday evening despite another quarter of ups and downs. Simply put, the leading AI chip maker has once again fallen victim to the curse of high expectations. However, this is not a concern for us, because Nvidia's fundamental fundamentals and long-term outlook look as healthy as ever. Revenue rose 94% year over year to a record $35.08 billion, easily surpassing the $33.16 billion the Street was looking for, according to estimates compiled by data provider LSEG. Adjusted earnings per share more than doubled to 81 cents, beating the consensus estimate of 75 cents, LSEG data showed. Current quarter guidance for revenue and gross margin were also above expectations, although clearly not as much as more bullish investors had hoped (let them sell, more for us). The stock fell about 2% in extended trading, to about $143 per share. Shares of Nvidia, the world's most valuable company, ended Wednesday's session up nearly 42% since their lows in early September. That marked the end of an unnecessarily sharp sell-off in response to its earnings report in late August. NVDA YTD The mountainous performance of Nvidia stock since the beginning of the year. Bottom line: Nvidia reported a great quarter on Wednesday — even if guidance for the current quarter was slightly below the highest expectations, weighing on the stock. It's hard to complain about the quarter hitting and lifting just because the hitting and lifting wasn't as great as some craved. Nvidia's earnings call made clear that we're in the early stages of an AI revolution that will fuel demand for Nvidia's market-leading chips through 2025 and likely well beyond. We reiterate our #1 rating and raise our price target on the stock to $165 per share, up from $150. Nvidia Blackwell's next-generation AI chip is in “full production,” Chief Financial Officer Colette Kress said. It is heading toward fiscal year 2026, which begins in earnest in February. Customers are hungry for chips. “We will ship Hopper and Blackwell (current generation) systems in the fourth quarter of fiscal year 2025 and beyond,” Chris said in written statements. “Both the Hopper and Blackwell systems face certain supply constraints, and demand for Blackwell is expected to exceed supply for several quarters in fiscal 2026.” Some investors may consider the supply crunch disappointing because it means leaving money on the table, at least until the middle of next year. But we don't worry. This is almost certainly a dynamic in which sales are pushed out, rather than lost entirely. For this reason, any material pullback in Nvidia stock driven by these constraints is a buy – and that's the advantage of being a long-term focused investor. Eventually, Blackwell's orders will be met, and given the company's efforts to update production lines on an annual basis, we'll actually hear about next-generation chips by the time supply catches up with demand. Not surprisingly, CEO Jensen Huang was asked about a recent media report that said a certain configuration of Blackwell chips was overheating. Huang was as dismissive as he could be, not that he was necessarily dodging the question. His answer made it seem like he didn't see the problem. He emphasized how advanced Blackwell is, both in the complex manufacturing process and in the process of actually installing them inside data centers. “This integration process (with customer-specific data centers) is something we've done for several generations now. We're very good at it,” Huang said. “But there's still a lot of engineering going on at this point….And as you can see from all the systems being put in place, Blackwell is in great shape.” Why we have it: Nvidia's high-performance graphics processing units (GPUs) are… The main driver behind the intelligence revolution Artificial, powering accelerated data centers that are being built rapidly around the world But this is more than just the hardware story With the Nvidia AI Enterprise service, Nvidia is building a potentially huge software business Competitors: Advanced Micro Devices and Intel's latest operation Buy: August 31, 2022 Start: March 2019 Huang once again raised another emerging concern in the investment community: Are the quality of AI models not improving as much as previously expected despite the additional computational power that is known in the… Technology industry as “expansion.” Think of it as hitting a ceiling where larger data centers with more GPUs don't result in much of an improvement in model capabilities — at least, not enough to justify all the extra spending on the latest and greatest hardware. Maybe one day this will happen, but according to Hwang, it doesn't look like it will be an issue anytime soon. When analyzing a particular AI model, Huang said there are three distinct stages in which it can become more advanced thanks to a larger amount of increasingly powerful chips: 1) the initial “pre-training” stage 2) the optimization process where fine-tuning is made 3) Real-world use known as heuristics. As Blackwell becomes more available and customers are able to benefit from advances in its performance over Hopper, there should be a marked improvement in model quality at every stage, the CEO said. He noted that the current generation of so-called core models — essentially large, general-purpose models — use about 100,000 Hopper chips. Now, as we start this next generation, we'll see models running on 100,000 Blackwell chips – and the expansion is still underway. Hwang's argument is that exceeding 100,000 Blackwell chips will lead to more capable models. The positive impact for Nvidia shareholders is that its customers are almost forced to buy more, or risk falling behind its competitors who do so. Huang also allayed concerns about a looming “digestion phase” following Hooper's move to Blackwell. This is when customers temporarily back off orders, enabling them to earn profits and realize a real return on the investments they have made in their existing computing infrastructure. It's a fair question to ask because Nvidia stock has historically taken a hit when its customers — like cloud computing providers — start uttering that phrase. “I think there will be no digestion until we modernize a trillion dollars' worth of data centers,” Huang said, adding: “If you just look at the world's data centers, the vast majority of them are built for a time when we've written applications by hand, and “Run them on central processing units (CPUs).” The update Hwang is referring to relates to GPU-centric data centers, geared toward a world of AI-written software. Nvidia also continues to see momentum in the AI space, said Chris, the CFO. Sovereign where Countries are embracing their own chip technology “for a new AI-powered industrial revolution” This is a growing market for Nvidia that is helping to expand its customer base beyond US tech giants like Microsoft, Meta Platforms and Amazon Add all these dynamics – supply constraints , the still-healthy expansion, the legacy positions, and the growing client pool — and it becomes abundantly clear that selling Nvidia stock based on its three-month guidance is the wrong approach. Investors will be better served by owning the stock for the long term. Instead of trying to trade in and out of every swing in the stock price. Guidance Taking a closer look at the guidance, Nvidia's fiscal Q4 outlook looks good versus consensus analyst estimates. However, investors have come to expect it – going ahead of expectations is the bottom line for this company. In the coming days, Wall Street is expected to debate whether the better-than-expected volume of forecasts justifies the stock returning to its all-time closing high of roughly $149 per share. It bears repeating: Determining whether management's expectations for the next three months, especially during a period of significant production, is not the way you maximize long-term upside. Instead, focusing on fundamental trends reveals a company with tremendous runway for growth ahead of it. Revenue was $37.5 billion, plus or minus 2%, ahead of the consensus estimate of $37.1 billion. This means that the annual growth rate is about 70%. Adjusted gross margins are expected to be 73.5%, plus or minus 50 basis points, slightly above estimates of 73.3%. Going forward, margins will remain about the same, and the forecast for fiscal fourth-quarter adjusted operating expenses of $3.4 billion appears to be slightly ahead of expectations of about $3.2 billion. Looking a little further, management said it's reasonable to assume Nvidia's gross margin ratio will return to the mid-70s by the back half of calendar year 2025. To be sure, executives made clear that will depend on the path of Blackwell's ramp and the company's selling mix. (Jim Cramer's Charitable Trust is long NVDA. See here for a complete list of stocks.) As a subscriber to the CNBC Investing Club with Jim Cramer, you will receive a trade alert before Jim makes a trade. Jim waits 45 minutes after a trade alert is sent before buying or selling a stock in his charitable fund's portfolio. If Jim talks about a stock on CNBC TV, he waits 72 hours after the trade alert is issued before executing the trade. The above Investment Club information is subject to our Terms and Conditions and Privacy Policy, as well as our Disclaimer. No obligation or fiduciary duty exists or is created by your receipt of any information provided in connection with the Investment Club. No specific results or profits are guaranteed.
Jensen Huang, co-founder and CEO of Nvidia Corp., holds the company's AI accelerator chips for data centers as he speaks during the Nvidia AI Summit Japan in Tokyo, Japan, on Wednesday, November 13, 2024.
Akio-kun | Bloomberg | Getty Images
Nvidia shares fell on Wednesday evening despite another quarter of ups and downs. Simply put, the leading AI chip maker has once again fallen victim to the curse of high expectations. However, this is not a concern for us, because Nvidia's fundamental fundamentals and long-term outlook look as healthy as ever.
