NVIDIA's Meteoric Rise: Analyzing Growth, Challenges, and Future Prospects

Nvidia Q2 Offers Several Clues To Sustainable Growth Ahead (NASDAQ:NVDA)

Nvidia's valuation is attractive for long-term investors, with strong financials, ongoing AI advancements, and significant future growth potential. Nvidia Corp's (NASDAQ:NVDA) double-digit percentage dip after Q2 2025 earnings were announced is an understandable one. Despite beating on both lines and more than doubling its quarterly revenues, the company's topline performance - and, in particular, forward guidance - clearly failed to meet investor expectations. The problem, as I see it, is not that Nvidia didn't put up solid numbers for the quarter. On the contrary, the company posted a near-102% topline growth rate between Q2 2023 and Q2 2024, but reported a much higher rate of 122% between Q2 2024 and Q2 2025, beating analyst estimates by nearly $1.3 billion. At the bottom line, too, the company reported adjusted earnings per share of 68 cents and beat estimates by 4 cents. Those aren't really numbers that would deserve a beat-down of the stock, so what's going on? Some analysts are saying that "The size of the beat this time was much smaller than we've been seeing." The chief argument seems to be that the company posted growth rates in the range of +200% over the past three quarters (on a YoY basis) but only showed a 122% increase in Q2. I find that argument to be faulty on many levels. Q2 2024 revenues grew by 102%, coming in at $13.5 billion and up from $6.7 billion in the year-ago quarter, so Nvidia is effectively showing much stronger growth this year with that 122% topline growth rate. We need to keep in mind that FY 2023 wasn't a great year for Nvidia. Revenue growth was flat from the year-ago period, barely $60 million higher. Indeed, Q4 2023 revenues of $6.05 billion were 21% down from Q4 2022. Go back one more quarter, and we see Q3 2023 revenues of $5.93 billion were down 17% from Q3 2022. So, we had a year of flat growth in FY 2023, but the revenue surge hadn't yet begun. In Q1 2024, Nvidia posted a -13% growth rate at the top at $7.19 billion, and it was not until Q2 2024, as we saw that the company started hitting positive growth rates - an in triple digits, no less. That's what Q2 2025 is comping against, and I don't see that as a growth rate decline; rather, I see it as the company having added nearly $7 billion to the top line between Q2 2023 and Q2 2024 for a 102% growth rate, but then adding $16.5 billion on top of that between Q2 2024 and Q2 2025. To me, it seems like the market (as well as analysts) is focusing too much on the percentage increase rather than the sheer dollar increase, which remains nothing less than superb. Adding $16.5 billion is definitely much harder than adding $7 billion, and many investors who have sold off since the latest earnings report seem to have missed that nuance. Moving to the Q3 guidance figure of $32.5 billion at the top, that equates to adding about $14.5 billion compared to the prior period. Again, the market seems to obsess about the figure 'only' representing an 80% growth rate, while the dollar value isn't being considered. Sure, you could look at that as a drop in growth rate, but we also need to remember that Nvidia's annualized revenue run-rate is now at $120 billion. It's not easy to keep adding the same percentage increases at that scale, but that doesn't stop Nvidia from adding real dollar volumes in a stable and sustainable manner. In short, while I'll concede that the apparently slower growth ahead is a disappointment to investors who were expecting the +200% cadence to continue, I don't believe it calls for a sell-off of this magnitude. As such, I think NVDA represents a compelling investment case even now. The correction is simply an opportunity for long-term investors to lower their cost basis. Growth is far from over, and that's what I plan to dive into with the rest of this article. My rating remains a Strong Buy after Q2 2025 earnings were released a little over a week ago. And this is why... One of the key concepts that investors need to understand is that Artificial Intelligence, or AI, is not a flash-in-the-pan technological development. It is a generational opportunity that has gained momentum over the past two years, but behind it are decades of experimentation, research, development, testing, and learning. One of the enablers of AI proliferation in recent years has been the ability to run these workloads on the cloud and leverage the power of scalability. More specific to my NVDA thesis is the fact that it is the growth in data center computer power that has altered the AI landscape and made it fertile enough for various industries to benefit significantly from. This is exactly where NVDA has a near-insurmountable lead over the competition - whether that's coming from Advanced Micro Devices (AMD) or any other chipmaker competing in the data center GPU space. As of 2023, NVDA shipped 3.76 million units for a mind-numbing 98% total shipment market share against a combined 90,000 units for AMD and Intel (INTC). Of course, that's when the data center GPU game was starving for competition, so with AMD aggressively pushing its MI300 agenda these past few quarters and looking at a Q4 release for the MI325 accelerator platform, that market share is likely to see some erosion as we navigate the final leg of H2 2024. That should actually be excellent news for Nvidia because it will keep management on its toes. I don't see it as a threat as much as an expansion of offerings in the space that addresses a range of price-point and performance requirements. My point, though, is that GPU shipments aren't likely to slow down their acceleration, and there are several indicators that point to this conclusion. As of Q2 2025, data center revenues topped $26.3 billion for a +150% increase YoY, and what I liked even more was the fact that customer concentration is spread out almost evenly between the hyperscale clients and other major customer cohorts - 45% and 50%, respectively. That's a nugget many investors might be missing because it implies that demand is spreading beyond cloud service providers and into on-prem and hybrid deployments. I expect to see that spread widening in favor of non-CSPs as we progress through FY 2025. This shift from a primarily CSP-heavy revenue base to a more expanded customer cohort mix isn't sudden. Indeed, management clued us in on this in the Q1 2025 press release from May 22 this year: Beyond cloud service providers, generative AI has expanded to consumer internet companies, and enterprise, sovereign AI, automotive and healthcare customers, creating multiple multibillion-dollar vertical markets. It also follows, logically, that since Blackwell products weren't widely available as of Q2 and likely won't in Q3, the bulk of revenue growth is still coming from the Hopper series of products, particularly the H200 platform that stepped up shipments in Q2. On that front, management had this to say: NVIDIA H200 platform began ramping in Q2, shipping to large CSPs, consumer Internet, and enterprise company. The NVIDIA H200 builds upon the strength of our Hopper architecture and offering over 40% more memory bandwidth compared to the H100. So, we should be seeing an increase in H200 shipments in Q3, which should tide the top line over nicely until Blackwell starts to pick up on shipments later this year. The company still expects a strong Q4 rollout for Blackwell, and the rumored delay is very likely linked to the improvements that were implemented to enhance Blackwell's production yields, in my opinion. These points were all validated (to a degree) by management: We executed a change to the Blackwell GPU mass to improve production yields. Blackwell production ramp is scheduled to begin in the fourth quarter and continue into fiscal year '26. In Q4, we expect to get several billion dollars in Blackwell revenue. Hopper shipments are expected to increase in the second half of fiscal 2025. So, what we're looking at for the two quarters ahead is a sharp increase in Hopper shipments and revenues for both Q3 and Q4, with a strong expectation of Q4 revenues getting a further boost from Blackwell sales. As such, I'm not really convinced by arguments that point to a gradual revenue decline for Nvidia. Yes, at this scale, growth on a percentage basis will naturally slow down, but from an absolute dollar value perspective, adding $14.5 billion in quarterly revenues on a YoY basis doesn't reflect a company that's seeing any kind of material decline in its largest segment. That's the first point of validation for my 'AI is not dead' claim. The second one, which I discuss in the next section, is even more interesting. Although Nvidia does offer air-cooled Blackwell compute platforms - B100 and B200 - with the HGX form factor, the Grace Blackwell accelerator needs liquid cooling, and data centers are already ramping up their infrastructure spending to include liquid cooling systems. There's clear evidence of that from multiple quarters. Vertiv Holdings (VRT), for example, which is collaborating with Nvidia "to build state-of-the-art liquid cooling solutions for next-gen NVIDIA accelerated data centres powered by GB200 NVL72 systems", in its Q2 earnings presentation, reported a $7 billion backlog that's up nearly 50% on a YoY basis and 11% on a sequential basis (from the $6.3 billion backlog reported at Q1 end). Of course, not all of that backlog is related to liquid cooling, but a strong supporting argument for growth in this industry niche comes from the fact that H2 2024 is expected to be the inflection point for the DCLC or data center liquid cooling market (emphasis mine): According to a recently published report from Dell'Oro Group, the trusted source for market information about the telecommunications, security, networks, and data center industries, the Data Center Liquid Cooling market has hit an inflection point, with mainstream adoption of liquid cooling starting in the second half of 2024. We forecast this to materialize over the next five years (2024-2028) in a market opportunity totaling more than $15 B. Beyond that, projections out to 2031 show a doubling of the DCLC market to $30 billion, which is a clear sign that DLC or direct-to-chip liquid cooling is here to stay. Of course, Nvidia does recognize that not all data centers are going to completely transition to liquid cooling, but one thing that CEO Huang said on the Q2 call really stuck in my mind: "the way to think about that is the next $1 trillion of the world's infrastructure will clearly be different than the last $1 trillion, and it will be vastly accelerated. With respect to the shape of our ramp, we offer multiple configurations of Blackwell. Blackwell comes in either a Blackwell classic, if you will, that uses the HGX form factor that we pioneered with Volta. And I think it was Volta. And so, we've been shipping the HGX form factor for some time. It is air-cooled. The Grace Blackwell is liquid-cooled. However, the number of data centers that want to go to liquid-cooled is quite significant. And the reason for that is because we can, in a liquid-cooled data center, in any data center -- power-limited data center, whatever size data center you choose, you could install and deploy anywhere from three to five times the AI throughput compared to the past. And so, liquid cooling is cheaper. Liquid cooling, our TCO is better, and liquid cooling allows you to have the benefit of this capability we call NVLink, which allows us to expand it to 72 Grace Blackwell packages, which has essentially 144 GPUs. The way I see it, the shift to liquid cooling is a tectonic one that will see TCO or total cost of ownership drop further as these technologies are increasingly adopted by data center operators. Right now, there's a capex hurdle to DCLC adoption, but the benefits of lower opex over extended periods of time greatly reduces the TCO for these systems. Over time, I believe we'll see much more momentum in this shift from air cooling to liquid cooling, and in turn, that supports a stronger case for Nvidia's liquid-cooled Blackwell products and the roadmap beyond. That brings me to my third point to support the 'AI is not dead' case. One of the biggest challenges to AI deployment is the inability to accurately assess the value it can bring to an organization. In a recent Gartner survey (results published in May 2024), nearly half of all respondents considered this to be their biggest hurdle. That's not very encouraging for AI as a whole, but the other aspect of this is that Gen AI or generative AI deployments are currently the main drivers of AI growth, and many companies are seeing tangible results from their Gen AI deployments. From an overview perspective, this is what was reported in early August: Early adopters report significant advantages, including improved consumer service (69%), streamlined workflows (54%), increased client satisfaction (48%), and better use of analytics (41%). Other tangible results reported are revenue increases (34%), improvements in ratings (40%), and boosts in team productivity (32%). Another report specific to payroll processing states that: integrating AI into the payroll system yielded a remarkable 60% increase in process efficiency -- a testament to tangible benefits: efficiency gains, cost savings, and error reduction. Yet another domain where AI excels and on the verge of showing tangible benefits is preventive healthcare, where algorithms are already doing the following: Diverse industries are benefitting or likely to soon benefit from AI deployments, not just Gen AI. In the face of this deluge of reports from every quarter, how can we keep claiming that AI is a bubble? The truth of the matter is, we've not even scratched the surface when it comes to AI capabilities, and I believe investors who sold off Nvidia on the basis of claims by some analysts about revenue declines, margin contraction, and other factors aren't taking the time to understand Nvidia's business and how it fits into the larger scheme of the AI revolution. As for my Strong Buy thesis, which I'm vehemently holding on to since I upgraded from Hold, it is based on some very simple ideas: We're still in the very early stages of AI development, and Nvidia is at the cutting edge of this generational transformation. However, what about investment-specific risks? With valuation levels still so high, can the stock move up further, or does it first need to grow into its current valuation before it can surge ahead? I believe that investors who couldn't get those questions answered to their satisfaction decided to take their money elsewhere. In a way, that's great for long-term investors. If you're still in the game, I'd urge you to consider the following: We're only at the beginning of the AI revolution, and the compute and networking needs of tomorrow's data centers need to be prepared for today, which is what Nvidia has been doing over the past several decades, ever since it invented the GPU in the late 1990s. My firm opinion is that there's still tremendous value in this company from a long-term capital appreciation perspective, and its meteoric rise over the past few years is not a flash-in-the-pan event that's one and done. By no means is this growth story over, and the sooner investors realize this, the better for their portfolios. Are there still material threats to the business? Absolutely! AMD is closing in on the GPU front, albeit very gradually, and in spurts. Although it may be years before it reaches the level of Nvidia, that gap is definitely closing. At the macro level, the current state of the economy leaves a lot to be desired. The high cost of debt could be hampering capex capabilities for thousands of potential customers of NVDA, so once the Fed starts its rate cut schedule, my assumption is that we should see material increases in capex allocations across multiple industries and segments. The only other major threat is a failure to execute on its plans for Blackwell. If those rumors I addressed in my last article were true and Blackwell doesn't start shipping in Q4, revenue growth may be impacted; worse, investors could start losing faith in management's ability to execute. On balance, I would say that this is still a very attractive investment opportunity, particularly after the macro, market, and results-related stock price corrections of the past two months. There seems to be some technical support at the current (pre-market) level of $105, but I'm not really concerned when I look at the long-term trajectory for the stock.

Nvidia: The Blackwell Delay And Its Consequences (NASDAQ:NVDA)

This idea was discussed in more depth with members of my private investing community, Rethink Technology. Learn More " Although Nvidia Corporation (NASDAQ:NVDA) reported fiscal 2025 Q2 results well above guidance, investors were disappointed with the guidance for fiscal Q3. This was mainly due to the delay in Nvidia's next-generation AI accelerator chip for the data center, dubbed Blackwell. Rumors have swirled around the delay, and some incorrect or misleading information about Blackwell has been reported. In this article, I take a closer look at the Blackwell delay and what it means for Nvidia's results in fiscal Q3 and Q4. On August 2, The Information released a report that Nvidia's Blackwell would be delayed by at least a quarter due to unspecified "design flaws." Normally, when one hears about the design flaws of a chip, logic design problems or bugs come to mind. But Nvidia management made clear that this was not the case during their fiscal Q2 results conference call. Instead, there was an issue with a mask that impacted chip yield, said CFO Colette Kress: Hopper demand is strong and Blackwell is widely sampling. We executed a change to the Blackwell GPU mask to improve production yields. Blackwell production ramp is scheduled to begin in the fourth quarter and continue into fiscal year '26. Presumably, the chip yield affects the number of good chips resulting from each silicon wafer fabricated by foundry partner TSMC (TSM). CEO Jensen Huang reiterated that there was nothing wrong with the functional design of the chip. The mask change did not change the functional logic of the chip. Developing the mask to implement a given layer of circuitry on an advanced chip has become an enormously complex process. This is because at the extreme ultraviolet wavelength of 13.5 nm, diffraction effects and optical distortion mean that the light pattern produced on the chip doesn't look like the mask. Optical physicists try to work backwards from the desired pattern to predict what the mask should look like, depending on the EUV machine and other factors. The process is computationally intensive, and TSMC has invested in an Nvidia supercomputer to perform what is called "computational lithography." The process isn't perfect, and the actual pattern produced by a mask may not match the computational lithography prediction. Since the issue is one of yield, this is probably what happened in the case of Blackwell, and why it's going to take several months to find a more optimal mask solution and begin mass production. Asa Fitch's article in the WSJ reported on the Blackwell delay: Nvidia hasn't detailed the nature of the issue. But analysts and industry executives say its engineering challenges stem mostly from the size of the Blackwell chips, which require a significant departure in design. I disagree with this interpretation. Nvidia has been making its flagship GPU accelerators (such as Hopper), at TSMC's reticle limit, for years. This sets the maximum physical size of a chip that TSMC can produce using EUV lithography machines produced by ASML Holding (ASML). Blackwell does consist of two such chips, but the process to make each is essentially unchanged from Hopper. Blackwell is fabbed, using basically the same TSMC N4 process as Hopper. Fitch continues: Instead of one big piece of silicon, Blackwell consists of two advanced new Nvidia processors and numerous memory components joined in a single, delicate mesh of silicon, metal and plastic. The manufacturing of each chip has to be close to perfect: serious defects in any one part can spell disaster, and with more components involved, there is a greater chance of that happening. Once again, I have a different interpretation. The fact that the package consists of two chips doesn't make the chips harder to make. Each chip already needs to be "close to perfect." The added complexity is in packaging, not in making the silicon. And it certainly doesn't have anything to do with the mask issue. The Ultra chips consist of two M-series Max chips joined edge-to-edge, as shown for the M1 Ultra: TSMC seems to be making a similar packaging approach available to Nvidia for Blackwell. This would indicate that the two Blackwell chips are identical and therefore no more difficult to make than Hopper. How Blackwell compares with AMD's MI300X, MI325X The strongest competition to Nvidia's data center AI accelerators is posed by Advanced Micro Devices (AMD) with its AMD Instinct line of GPU accelerators. The Blackwell approach is certainly less complex than what AMD uses for its flagship data center accelerator, the MI300X, which uses multiple chiplets and interposers: The MI300X has been hailed as an "Nvidia Killer," yet when AMD launched it, AMD could only claim performance parity in AI training to the Hopper H100: However, Nvidia claims a huge performance improvement for Blackwell over Hopper H100 in both training and inference: Nvidia has also published some early results on MLCommons, which tabulates AI benchmarks for systems that run MLPerf tests. In the llama2-70B inference benchmark, the Blackwell B200 bested the MI300X by a factor of 3.67. In calendar Q4, AMD is expected to release the MI325X, which is basically identical to the MI300X but with more memory and higher memory bandwidth. The memory improvements will help performance, but will probably not close the gap with Blackwell. The MI300 series chips are powerful accelerators, but they fall short of Nvidia in AI performance because they were never really meant for AI. Originally, the MI300A was intended to power the El Capitan supercomputer of Lawrence Livermore. This was what AMD refers to as an "APU," an SOC with CPU and GPU cores: Nvidia didn't specify how much wastage there was due to the Blackwell mask problem, but apparently some wafers already bought and paid for from TSMC had to be scrapped. This was disclosed as a contributor to a sequential decline of gross margin from 78.35% in fiscal Q1 to 75.15% in Q2. With production Blackwell chips not scheduled to ship until fiscal Q4, Q3 guidance was muted but still very respectable with 79% y/y revenue growth and net income y/y growth of 84%. But it just couldn't compare with the huge growth rates of quarters past. And it couldn't possibly. Fiscal 2024 Q3 was when generative AI really took off, and company revenue grew by over 200%. This "law of large numbers" effect merely means that percent changes are going to get smaller, even if the magnitude of the revenue change stays the same. The Data Center Delta is the y/y absolute difference in revenue. In fiscal 2025 Q1, even though the delta had continued to get larger, the percent change leveled off. The y/y percent change plunged in fiscal Q2, even though the delta decreased only slightly from fiscal Q1. The expectation that revenue growth should be always proportionate to revenue is probably not realistic. Allow me to suggest a better way to look at Nvidia's Data Center revenue trajectory. Since Q2, that trajectory has been nearly a straight line: The slope of the line can be approximated as the difference between current quarter revenue and the previous quarter, divided by 1 quarter of time. In effect, the slope is just given by the sequential revenue difference or delta, shown plotted in yellow. Where the line is straight or nearly so, the slope is constant. The slope is greatest in fiscal 2024 Q2, when the AI boom began for the Data Center segment. But the constant slope in the following quarters implies that the revenue growth was sustainable. Then came the Blackwell delay, and the midpoint of revenue guidance implies a decreased rate of growth in the Data Center, as shown below: If the slope of the line were expected to continue to fall, this would be very concerning. But there's no reason to expect that it will. Nvidia expects demand for Hopper H200 to continue to grow into Q4, while Blackwell adds additional revenue to the data center. Kress said: In Q4, we expect to ship several billion dollars in Blackwell revenue. Hopper shipments are expected to increase in the second half of fiscal 2025. Hopper supply and availability have improved. Demand for Blackwell platforms is well above supply, and we expect this to continue into next year. This is reflected in my projection for fiscal Q4 Data Center revenue and the recovery of the slope line. The production snag that Blackwell hit was nothing unusual in advanced semiconductor fabrication and would probably have gone unnoticed were it not for the accelerated schedule for Blackwell's release. Previous releases have been far more leisurely. The delay says nothing about the fundamental viability or producibility of the chip. The packaging technology used for Blackwell is ambitious but by no means new. And the chip represents another quantum leap in AI performance and computational efficiency. The AI industry needs that efficiency to keep from draining away all the world's electrical generating capacity. We're in the early stages of a roughly $2 trillion revamp of the world's data centers. At Nvidia's 2024 GPU Technology conference in March, during a Q&A with analysts, Huang estimated that the data center infrastructure spending rate was about $250 billion per year. Depending on how long the spending rate is sustained, that would be about $1-2 trillion in spending over the next ten years. Also during Q&A, Vivek Arya of Bank of America bounded the total addressable market as being in the $1-2 trillion range. I believe that with the growth of sovereign AI, where countries invest in domestic AI infrastructure, that we'll be at the upper bound of that range by the end of the decade. This revamp is being pursued not merely for the sake of AI, but for purposes of greater efficiency as data center operators turn to GPU acceleration for a wide variety of workloads besides AI. These include big data analytics, video and game streaming, metaverse applications, etc. One of the advantages of GPU accelerators over special purpose ASICs is versatility, the ability to do other useful things besides AI. This is why I consider AMD the strongest competitor. Do I think that Nvidia will scoop up all the $2 trillion data center opportunity? No, I don't. In addition to AMD, Nvidia will face competition from in-house developed AI accelerators such as Google's (GOOG) (GOOGL) TPU, and chips developed by Amazon (AMZN), and even, as was recently disclosed at WWDC, by Apple. I'm currently making a conservative assumption that Nvidia's Data Center cumulative revenue from fiscal 2025 through 2029 will be $955 billion, or less than half of the projected opportunity. But even then, I expect Nvidia to be the dominant player in the space. I remain long Nvidia stock and rate it a Buy.

Is Nvidia Overvalued? The $3 Trillion Question

Last week Nvidia released its quarterly earnings to much fanfare (including this Nvidia earnings watch party in NYC). However, despite beating Wall Street's expectations yet again, the stock declined as much as 5% in after-hours trading. For some context, Nvidia became the 2nd most valuable company in the world (at ~$3 trillion) because of the AI revolution. Following the release of ChatGPT in November 2022, the demand for AI/LLMs (large language models) exploded. This meant more data centers and more computing power. The backbone of all of this additional compute comes from GPUs (graphical processing units). That's what Nvidia produces and it produces them better than anyone else. As a result, the demand for their GPUs went through the roof. Today, many of the largest tech companies are spending a sizable portion of their total capital expenditures on Nvidia's GPUs. For example, Microsoft spends nearly 40 percent of their total capital expenditures on Nvidia, up from less than 10% about a year ago. You can see just how much money many of the tech giants are sending to Nvidia in this chart from Yahoo Finance: This explains why Nvidia has 80% to 95% market share in AI computing. But it's not just their total market share that matters, but their growth. As this chart from Yahoo Finance illustrates, Nvidia's data center revenue is up 6x over the prior five quarters: When you combine this incredible growth in revenue with their near monopoly on GPUs, you get a stock that is up 150% YTD. But as impressive as Nvidia's growth has been, I'm here to tell you that its valuation is out of control. To provide some context for this statement, let's look at the Price-to-Sales (P/S) ratio of Nvidia today compared with Microsoft's P/S ratio during the DotCom bubble. I've aligned these two measures such that the peak of each P/S ratio occurs at the same time (i.e. at the 2,500th trading day) in the chart below: Nvidia is trading at a higher P/S multiple today than Microsoft was during the peak of the DotCom bubble! And do you know what happened to Microsoft after its P/S ratio peaked? Its stock got cut in half and didn't fully recover for 14.5 years. From January 2000 to July 2014, Microsoft returned 0% while the S&P 500 went up 77% on a total return basis: Of course, I am not implying that the same exact thing will happen to Nvidia, but you can see why it's concerning. I generally don't like relying on valuation metrics, however, the P/S ratio is one I monitor in the extremes. The reason why is because it's the purest valuation metric out there. You can't manipulate it with accounting tricks like you can earnings (P/E) or book value (P/B). Outside of outright fraud, sales are difficult to fake. As a result, this metric can be quite reliable during frothy periods. The last time I wrote about general stock market froth in November 2021 I relied on the P/S ratio. This is what it looked like at the time: Within a few months of that post, many of these high P/S stocks were crashing. If we were to re-run this chart through today, it would look something like this: Note that the big spike in early 2022 occurred after I published my "This Will Not Last" warning in November 2021. Even when you know something is off, getting the timing right is incredibly difficult. Either way, you can see that there are some frothy parallels between now and 2021. It's not just me who has pointed this out either. Apollo Global Management created this chart illustrating that the 12-month forward P/E ratio of the top 10 companies in the S&P 500 is higher today than it was during the DotCom bubble: I don't want to draw too many comparisons to the DotCom bubble because this seems quite different. Though valuations as a whole are slightly elevated, this post isn't meant to be a "This Will Not Last" Part II. If there is any market frothiness today, that frothiness is incredibly concentrated. For example, in the chart above showing the market capitalization of companies with a P/S ratio > 20, nearly 66% of that market capitalization today is one company -- Nvidia. That wasn't true in 2021. Yes, there was definitely concentration in the tech sector, but not like this. So, if you own lots of Nvidia stock, I will say the same thing that I said to Tesla shareholders back in early 2021 -- just take the money. You've won the game. You've hit a lotto that you are unlikely to hit again. Take at least some of your gains and diversify. Go on vacation. Enjoy life. Don't get me wrong, if you listen to me you are likely to regret it in the coming months. However, I doubt you will regret it in the coming years. Those Tesla shareholders that heeded my warning on January 12, 2021 and sold their shares would've missed out on a 45% gain through November 2021. However, those same investors also wouldn't be down 27% on their Tesla shares through today either. Anyone who rotated out of Tesla and into the S&P 500 on January 12 2021, would be up 55% instead: Of course, maybe this time is different. Maybe Nvidia's near monopoly won't be challenged by regulators. Maybe the efforts of competitors to develop their own chips will all fail. Maybe the tech giants will be happy to keep sending 10%+ of their capital expenditures to Nvidia indefinitely. But, I doubt it. Not only does Nvidia have to deal with its own challenges from competitors and regulators, but with the shortcomings of AI itself. Yes, the AI/LLM revolution is real, but there's still a lot to be desired. It reminds me of this quote I read from an anonymous Twitter user about AI: "if my dog could summarize the newspaper to me every morning but got 20% of the stories wrong i would be fascinated by the dog but would still read the newspaper the normal way." For the record, I am a big fan of LLMs. I wouldn't have been able to build my S&P 500 historical return calculator without them. However, like any technology, they have serious limitations. Just consider this simple example (and why it happens): That's going to take over the world? In all seriousness, I bring up the flaws with AI because of how they might affect Nvidia downstream. If the demand for AI declines, so will the demand for Nvidia's GPUs. While this seems unlikely right now, the future is always uncertain. With that being said, I don't know what Nvidia's stock will do over the next month, the next year, or the next decade. I don't know a lot about AI or GPUs and I don't claim to. But I do know market history. And the last few times the data told me what it's telling me now, it didn't end well for those investors. Don't trade maybes for certainties. They're rarely worth the risk. Thank you for reading. If you liked this post, consider signing up for my newsletter. Market News and Data brought to you by Benzinga APIs

Where Will Nvidia Stock Be in 5 Years? | The Motley Fool

The chipmaker has delivered phenomenal returns to investors in the past five years, and it won't be surprising to see it remain a solid pick for the future as well. The last five years have been incredible for Nvidia (NVDA -4.08%) investors as shares of the semiconductor specialist have taken off in stunning fashion thanks to the terrific growth in its revenue and earnings, and a couple of solid catalysts in the form of video gaming and artificial intelligence (AI). An investment of just $100 in shares of Nvidia five years ago is now worth a whopping $2,850. The stock's gains of 2,750% during this period have crushed the S&P 500 index's jump of 93% in the past five years. But now, Nvidia is the world's third-largest company with a market capitalization of just over $2.9 trillion. Expecting the stock to jump another 25 times to 30 times from current levels over the next five years doesn't seem logical as its market cap would then be worth a whopping $80 trillion. The global economy, for comparison, was worth an estimated $105 trillion last year. However, it is worth noting that Nvidia still has solid-growth drivers in the bag that could allow it to sustain healthy levels of growth over the next five years. In this article, we will take a closer look at those catalysts and check how much upside this semiconductor stock could deliver over the next five years. Nvidia released fiscal 2025 second-quarter results (for the three months ended July 28) on Aug. 28. Its numbers turned out to be better than expected, with revenue jumping 122% year over year to $30 billion and non-GAAP earnings per share increasing by an impressive 152% from the same quarter last year to $0.68 per share. Wall Street analysts would have settled for $0.65 per share in earnings on $28.7 billion in revenue from the company. This is not where the good news ended as Nvidia is expecting its fiscal Q3 revenue to land at $32.5 billion, ahead of the $31.7 billion consensus estimate. That would translate into year-over-year growth of 80%. Still, the stock has been slipping despite reporting such phenomenal growth and delivering stronger-than-expected guidance. This slight pullback, however, is an opportunity for investors to buy the stock as a closer look at its latest quarterly results indicate that it is witnessing solid growth across all of its businesses. The data-center segment, which grew a whopping 154% year over year and clocked record revenue of $26.3 billion, is benefiting big time from the robust demand for Nvidia's graphics processing units (GPUs), which are being deployed to train and deploy AI models. What's worth noting here is that customers have continued to purchase Nvidia's Hopper architecture-based GPUs even though the company is set to begin the production ramp of its next-generation Blackwell chips in the next quarter. As pointed out by CFO Colette Kress on the latest earnings conference call: Blackwell production ramp is scheduled to begin in the fourth quarter and continue into fiscal year '26. In Q4, we expect to get several billion dollars in Blackwell revenue. Hopper shipments are expected to increase in the second half of fiscal 2025. Hopper supply and availability have improved. Demand for Blackwell platforms is well above supply, and we expect this to continue into next year. More importantly, Nvidia believes that the next-generation AI models are likely to require 10 to 20 times more computing power, suggesting that the demand for its data-center GPUs could keep growing in the long run. According to one estimate, the AI chip market could generate $311 billion in annual revenue in 2029 as compared to this year's estimate of $123 billion. Nvidia generated almost $49 billion in data-center revenue in the first half of its current fiscal year, indicating that it could end fiscal 2025 (which coincides with 11 months of calendar 2024) with almost $100 billion in revenue from this segment. Based on the $123 billion estimate of the overall AI chip market, Nvidia's share of this market could stand at around 80% by the end of this year. Even if Nvidia's AI chip share falls over the next five years to even 70% thanks to the emerging competition, it could still generate $217 billion in annual data-center revenue after five years (based on the $311 billion size of the overall market). In simpler words, Nvidia's data-center revenue may double over the next five years. Throw in the fact that the company's other businesses are also delivering healthy growth, it won't be surprising to see Nvidia becoming a bigger company over the next five years. For instance, the company's revenue from the gaming and AI personal computer (PC) segment was up 18% year over year to $2.6 billion. Gaming has been Nvidia's bread and butter over the years, but now it contributes less than 10% to its top line. However, investors would do well to look at the bigger picture as the demand for AI PCs is set to boom over the next five years, paving the way for Nvidia to generate substantial revenue from the gaming and PC segment. Meanwhile, the company's professional-visualization revenue increased 45% year over year to $427 million thanks to the growing demand for the company's digital-twin solutions. As Kress remarked on the latest earnings call: The world's largest electronics manufacturer, Foxconn, is using NVIDIA Omniverse to power digital twins of the physical plants that produce NVIDIA Blackwell systems. And several large global enterprises, including Mercedes-Benz, signed multiyear contracts for NVIDIA Omniverse Cloud to build industrial digital twins of factories. Again, this is another lucrative growth opportunity for Nvidia as the digital-twin market could generate $131 billion in revenue in 2029 as compared to $26 billion this year, according to Mordor Intelligence. In all, there are multiple reasons why Nvidia could turn out to be a solid investment over the next five years. Nvidia's healthy prospects are the reason why analysts have ramped up their revenue-growth expectations for the company's current fiscal year and beyond. Similarly, its bottom-line growth expectations have also seen a bump. The good part is that Nvidia is expected to clock annual-earnings growth of 52% for the next five years. Based on the company's fiscal 2025 earnings-per-share estimate of $2.84, its bottom line could jump to $23 per share after five years. The stock currently has a forward-earnings multiple of 45. However, even if it trades at a discounted 29 times forward earnings after five years, in line with the Nasdaq-100 index's forward-earnings multiple (using the index as a proxy for tech stocks), its stock price could hit $667. That would be a jump of over 5 times from current levels, suggesting that this AI stock could continue to make investors richer in the next five years.

The Power Of Expectations: Nvidia's Earnings And The Market Reaction

The company's performance over the last three quarters, in particular, have created expectations that no company can meet. Last Wednesday (August 28), the market waited with bated breath for Nvidia's (NASDAQ:NVDA, NEOE:NVDA:CA) earning call, scheduled for after the market closed. That call, at first sight, contained exceptionally good news, with revenues and earnings coming in at stratospheric levels and above expectations, but the stock fell in the aftermath, down 8% in Thursday's trading. That drop of more than $200 billion in market capitalization in response to what looked like good news, at least on the surface, puzzled market observers, though, as is their wont, they had found a reason by day end. This dance between companies and investors, playing out in expected and actual earnings, is a feature of every earnings season, especially so in the United States, and it has always fascinated me. In this post, I will use the Nvidia earnings release to examine what news, if any, is contained in earnings reports, and how traders and investors use that news to reframe their thinking about stocks. When I was first exposed to financial markets in a classroom, I was taught about information being delivered to markets, where that information is processed and converted into prices. I was fascinated by the process, an interplay of accounting, finance and psychology, and it was the subject of my doctoral thesis, on how distortions in information delivery (delays, lies, mistakes) affects stock returns. In the real world, that fascination has led me to pay attention to earnings reports, which while overplayed, remain the primary mechanism for companies to convey information about their performance and prospects to markets. Publicly traded companies have had disclosure requirements for much of their existence, but those requirements have become formalized and more extensive over time, partly in response to investor demands for more information and partly to even the playing field between institutional and individual investors. In the aftermath of the great depression, the Securities Exchange Commission was created as part of the Securities Exchange Act, in 1934, and that act also required any company issuing securities under that act, i.e., all publicly traded firms, make annual filings (10Ks) and quarterly filings (10Qs), that would be accessible to investors. The act also specifies that these filings be made in a timely manner, with a 1946 stipulation the annual filings being made within 90 days of the fiscal year-end, and the quarterly reports within 45 calendar days of the quarter-end. With technology speeding up the filing process, a 2002 rule changed those requirements to 60 days, for annual reports, and 40 days for quarterly reports, for companies with market capitalizations exceeding $700 million. While there are some companies that test out these limits, most companies file well within these deadlines, often within a couple of weeks of the year or quarter ending, and many of them file their reports on about the same date every year. If you couple the timing regularity in company filings with the fact that almost 65% of listed companies have fiscal years that coincide with calendar years, it should come as no surprise that earnings reports tend to get bunched up at certain times of the year (mid-January, mid-April, mid-July and mid-October), creating "earnings seasons". That said, there are quite a few companies, many of them high-profile, that preserve quirky fiscal years, and since Nvidia's earnings report triggered this post, it is worth noting that Nvidia has a fiscal year that ends on January 31 of each year, with quarters ending on April 30, July 31 and October 31. In fact, the Nvidia earnings report on August 28 covered the second quarter of this fiscal year (which is Nvidia's 2025 fiscal year). While corporate earnings reports are delivered once a quarter, the work of anticipating what you expect these reports to contain, especially in terms of earnings per share, starts almost immediately after the previous earnings report is delivered. In fact, a significant portion of sell side equity research is dedicated to this activity, with revisions made to the expected earnings, as you get closer and closer to the next earnings report. In making their earnings judgments and revisions, analysts draw on many sources, including: The earnings expectations for individual companies, from sell side equity research analysts are publicly accessible, giving us a window on trend lines. Nvidia is one of the most widely followed companies in the world, and most of the seventy plus analysts who publicly follow the firm play the estimation game, leading into the earnings reports. Ahead of the most recent second-quarter earnings report, the analyst consensus was that the company would report revenues of $28.42 billion for the quarter, and fully diluted earnings per share of 64 cents; in the 30 days leading into the report, the earnings estimates had drifted up mildly (about 0.1%), with the delay in the Blackwell (NVidia's new AI chip) talked about but not expected to affect revenue growth near term. It is worth noting that not all analysts tracking the stock forecast every metric, and that there was disagreement among them, which is also captured in the range on the estimates; on earnings per share, for instance, the estimates ranged from 60 to 68 cents, and on revenues, from $26 to $30 billion. The pre-game show is not restricted to analysts and investors, and markets partake in the expectations game in two ways. While volatility tends to increase just ahead of earnings reports, the surge in volatility ahead of the second quarter earnings for Nvidia was unusually large, a reflection of the disagreement among investors about how the earnings report would play out in the market. Put simply, even before Nvidia reported earnings on August 28, markets were indicating more unease about both the contents of the report and the market reaction to the report, than they were with prior earnings releases. Given the lead-in to earnings reports, what exactly do they contain as news? The SEC strictures that companies disclose both annual and quarterly results have been buffered by accounting requirements on what those disclosures should contain. In the United States, at least, quarterly reports contain almost all of the relevant information that is included in annual reports, and both have suffered from the disclosure bloat that I called attention to in my post on disclosure diarrhea. Nvidia's second quarter earnings report, weighing in at 80 pages, was shorter than its annual report, which ran 96 pages, and both are less bloated than the filings of other large market-cap companies. The centerpieces of the earnings report, not surprisingly, are the financial statements, as operating numbers are compared to expectations, and Nvidia's second quarter numbers, at least at first sight, are dazzling: The company's astonishing run of the last few years continues, as its revenues, powered by AI chip sales, more than doubled over the same quarter last year, and profit margins came in at stratospheric levels. The problem, though, is that the company's performance over the last three quarters, in particular, have created expectations that no company can meet. While it is just one quarter, there are clear signs of more slowing to come, as scaling will continue to push revenue growth down, the unit economics will be pressured as chip manufacturers (TSMC) push for a larger slice and operating margins will decrease, as competition increases. Over the last two decades, companies have supplemented the financial reports with guidance on key metrics, particularly revenues, margins and earnings, in future quarters. That guidance has two objectives, with the first directed at investors, with the intent of providing information, and the second at analysts, to frame expectations for the next quarter. As a company that has played the expectations game well, it should come as no surprise that Nvidia provided guidance for future quarters in its second quarter report, and here too, there were reminders that comparisons would get more challenging in future quarters, as they predicted that revenue growth rates would come back to earth, and that margins would, at best, level off or perhaps even decline. Finally, in an overlooked news story, Nvidia announced that it would have authorized $50 billion in buybacks, over an unspecified time frame. While that cash return is not surprising for a company that has become a profit machine, it is at odds with the story that some investors were pricing into the stock of a company with almost unlimited growth opportunities in an immense new market (AI). Just as Meta and Alphabet's dividend initiations signaled that they were approaching middle age, Nvidia's buyback announcement may be signaling that the company is entering a new phase in the life cycle, intentionally or by accident. The final piece of the earning release story, and the one that gets the most news attention, is the market reaction to the earnings reports. There is evidence in market history that earnings reports affect stock prices, with the direction of the effect depending on how actual earnings measure up to expectations. While there have been dozens of academic papers that focus on market reactions to earnings reports, their findings can be captured in a composite graph that classifies earnings reports into deciles, based upon the earnings surprise, defined as the difference between actual and predicted earnings: As you can see, positive surprises cause stock prices to increase, whereas negative surprises lead to price drops, on the announcement date, but there is drift both before and after surprises in the same direction. The former (prices drifting up before positive and down before negative surprises) is consistent with the notion that information about earnings surprises leaks to markets in the days before the report, but the latter (prices continuing to drift up after positive or down after negative surprises) indicates a slow-learning market that can perhaps be exploited to earn excess returns. Breaking down the findings on earnings reports, there seems to be evidence that the that the earnings surprise effect has moderated over time, perhaps because there are more pathways for information to get to markets. Nvidia is not only one of the most widely followed and talked about stocks in the market, but one that has learned to play the expectations game well, insofar as it seems to find a way to beat them consistently, as can be seen in the following table, which looks at their earnings surprises over the last 5 years: Barring two quarters in 2022, Nvidia has managed to beat expectations on earnings per share every quarter for the last five years. There are two interpretations of these results, and there is truth in both of them. The first is that Nvidia, as with many other technology companies, has enough discretion in both its expenditures (especially in R&D) and in its revenue recognition, that it can use it to beat what analysts expect. The second is that the speed with which the demand for AI chips has grown has surprised everyone in the space (company, analysts, investors) and that the results reflect the undershooting on forecasts. Focusing specifically on the 2025 second quarter, Nvidia beat analyst expectations, delivering earnings per share of 68 cents (above the 64 cents forecast) and revenues of $30 billion (again higher than the $28.4 billion forecast), but the percentage by which it beat expectations was smaller than in the most recent quarters. That may sound like nitpicking, but the expectations game is an insidious one, where investors move the goal posts constantly, and more so, if you have been successful in the past. On August 28, after the earnings report, Nvidia saw share prices drop by 8% and not only did that loss persist through the next trading day, the stock has continued to lose ground, and was trading at $106 at the start of trading on September 6, 2028. So what do you learn from earnings reports that may cause you to reassess what a stock is worth? The answer will depend upon whether you consider yourself more of a trader or primarily an investor. If that distinction is lost on you, I will start this section by drawing the contrast between the two approaches, and what each approach is looking for in an earnings report. Value versus Price At the risk of revisiting a theme that I have used many times before, there are key differences in philosophy and approach between valuing an asset and pricing it. The difference between investing and trading stems from this distinction between value and price. Investing is about valuing an asset, buying it at a price less than value and hoping that the gap will close, whereas trading is almost entirely a pricing game, buying at a low price and selling at a higher one, taking advantage of momentum or mood shifts. Given the very different perspectives the two groups bring to markets, it should come as no surprise that what traders look for in an earnings report is very different from what investors see in that same earnings report. If prices are driven by mood and momentum, it should come as no surprise that what traders are looking for in an earnings report are clues about how whether the prevailing mood and momentum will prevail or shift. It follows that traders tend to focus on the earnings per share surprises, since its centrality to the report makes it more likely to be a momentum-driver. In addition, traders are also swayed more by the theater around how earnings news gets delivered, as evidenced, for instance, by the negative reaction to a recent earnings report from Tesla, where Elon Musk sounded downbeat, during the earnings call. Finally, there is a significant feedback loop, in pricing, where the initial reaction to an earnings report, either online or in the after market, can affect subsequent reaction. As a trader, you may learn more about how an earnings report will play out by watching social media and market reaction to it than by poring over the financial statements. For Nvidia, the second quarter report contained good news, if good is defined as beating expectations, but the earnings beat was lower than in prior quarters. Coupled with sober guidance and a concern the stock had gone up too much and too fast, as its market cap had increased from less than half a trillion to three trillion over the course of two years, the stage was set for a mood and momentum shift, and the trading since the earnings release indicates that it has happened. Note, though, that this does not mean that something else could not cause the momentum to shift back, but before you, as an Nvidia manager or shareholder, are tempted to complain about the vagaries of momentum, recognize that for much of the last two years, no stock has benefited more from momentum than Nvidia. For investors, the takeaways from earnings reports should be very different. If value comes from key value inputs (revenues growth, profitability, reinvestment and risk), and these value inputs themselves come from your company narrative, as an investor, you are looking at the earnings reports to see if there is information in them that would change your core narrative for the company. Thus, an earnings report can have a significant effect on value, if it significantly changes the growth, profitability or risk parts of your company's story, even though the company's bottom line (earnings per share) might have come in at expectations. Here are a few examples: In short, if you are an investor, the most interesting components of the report are not in the proverbial bottom line, i.e., whether earnings per share came in below or above expectations, but in the details. Finally, as investors, you may be interested in how earnings reports change market mood, usually a trading focus, because that mood change can operate as a catalyst that causes the price-value gap to close, enriching you in the process. The figure below summarizes this section, by first contrasting the value and pricing processes, and then looking at how earnings releases can have different meanings to different market participants. As in other aspects of the market, it should therefore come as no surprise that the same earnings report can have different consequences for different market participants, and it is also possible that what is good news for one group (traders) may be bad news for another group (investors). My trading skills are limited, and that I am incapable of playing the momentum game with any success. Consequently, I am not qualified to weigh in on the debate on whether the momentum shift on Nvidia is temporary or long term, but I will use the Nvidia second quarter earnings report as an opportunity to revisit my Nvidia story and to deliver a September 2024 valuation for the company. My intrinsic valuation models are parsimonious, built around revenue growth, profit margins and reinvestment, and I used the second quarter earnings report to review my story (and inputs) on each one: With these input changes in place, I revalued Nvidia at the start of September 2024, breaking its revenues, earnings and cash flows down into three businesses: an AI chip business that remains its central growth opportunity, and one in which it has a significant lead on the competition, an auto chip business where it is a small player in a small game, but one where there is potential coming from demand for more powerful chips in cars, and the rest, including its existing business in crypto and gaming, where growth and margins are solid, but unlikely to move dramatically. While traders may be disappointed with Nvidia's earnings release, and wish it could keep its current pace going, I think it is both unrealistic and dangerous to expect it to do so. In fact, one reason that my story for Nvidia has become more expansive, relative to my assessment in June 2023, is that the speed with which AI architecture is being put in place is allowing the total market to grow at a rate far faster than I had forecast last year. In short, relative to where I was about a year ago, the last four earnings reports from the company indicate that the company can scale up more than I thought it could, has higher and more sustainable margins than I predicted and is perhaps less exposed to the cycles that the chip business has historically been victimized by. With those changes in place, my value per share for Nvidia in is about $87, still about 22% below the stock price of $106 that the stock was trading at on September 5, 2024, a significant difference but one that is far smaller than the divergence that I noted last year. As always, the normal caveats apply. The first is that I value companies for myself, and while my valuations drive my decisions to buy or sell stocks, they should not determine your choices. That is why my Nvidia valuation spreadsheet is available not just for download, but for modification, to allow you to tell your own story for Nvidia, yielding a different value and decision. The second is that this is a tool for investors, not traders, and if you are playing the trading game, you will have to reframe the analysis and think in terms of mood and momentum. Looking back, I am at peace with the decision made in the summer of 2023 to shed half my Nvidia shares, and hold on to half. While I left money on the table, with the half that I sold, I have been richly compensated for holding on to the other half. I am going to count that as a win and move on! Editor's Note: The summary bullets for this article were chosen by Seeking Alpha editors.

NVDL: Pain Awaits (NASDAQ:NVDL)

Insider stock sales and a DOJ antitrust subpoena further reinforce the view that Nvidia's meteoric rise may be nearing its end. Before we get too far into this, leveraged ETFs are generally very risky financial products. I've covered a few of them previously for Seeking Alpha here, here, and here. The way these types of funds typically work is they rebalance daily to optimize returns. This works quite well when shareholders time entries at the beginning of a bullish run in the underlying stock or index. Such has indeed been the case with the GraniteShares 2x Long NVDA Daily ETF (NASDAQ:NVDL). In this article, we'll look at fund details, the impressive performance of NVDL since fund inception, and I'll make my case why I think the party is probably over. As one has likely deduced already from the name of the fund, NVDL exists solely to produce twice the daily returns of Nvidia Corporation (NVDA) stock. For instance, on a day when NVDA shares rise by 5%, NVDL should move up by about 10%. The same is true on the way down. If NVDA goes down by 5%, NVDL should decline by 10%. Since the fund is rebalanced daily to optimize for double returns, NVDL's price should theoretically decay over time if NVDA experiences any sustained period of chop or decline. Thus, NVDL should really only be held when NVDA is in a strong uptrend and for short periods of time in a tactical strategy. Notably, NVDL is not the only product on the market that aims to produce 2x returns of NVDA stock. Other funds offering the same strategy include the Direxion Daily NVDA Bull 2x Shares (NVDU) and the T-Rex 2x Long NVIDIA Daily Target ETF (NVDX): Source: Seeking Alpha. Due to its 1.15% expense ratio, NVDL is slightly pricier for those who hold for an entire year (I don't recommend holding leveraged ETFs this long). Though when we're talking about a fund that doubles the daily return of NVDA, I suspect a few bps likely doesn't matter all that much to shareholders. Perhaps what does matter more is the track record. GraniteShares' fund is by far the largest at $5 billion assets under management ("AUM") and was launched before each of the other two funds. Comparing each of the funds over a year to date basis perhaps gives the most fair picture because neither NVDU nor NVDX existed a year ago: Here we see three funds that have all produced terrific year to date gains, though the total returns are actually quite different. NVDL is right in the middle at 207% year to date, while NVDU is well under that at 167% and NVDX is in the lead at 237%. Compared to NVDA, only NVDX has delivered on a true 2x goal since the beginning of 2024, though NVDL is relatively close. Again, NVDL has been around longer than the other two funds, and so looking at the fund's total performance may be helpful: Since the fund's inception near the end of 2022, the total return for Day 1 shareholders has been over 1,100%. This is against NVDA's 500% total return over the same period of time. Thus, NVDL has actually achieved more than double the 2x daily goal of the fund over the last 20 months. In my experience both covering and utilizing these types of funds for trading, this is not a typical story, and it speaks to the intensity of the NVDA bull run over the last couple of years. It's been incredibly impressive, to say the least. However, I believe that this has generally been an outlier in the world of 2x leveraged ETFs, and it's likely going to come to an end. I'll get ahead of the disclosure at the bottom of the page to be as transparent as possible. I have a synthetic short position on NVDA through a long position in the Tradr 1.25X Short NVDA Daily ETF (NVDS). If NVDA goes down, I will make money. I've taken this position for several reasons that include but are not limited to the following; We'll start first with the sentiment signals. Regarding small sample sizes and anecdotal evidence as justification for a broader viewpoint, I'm just seeing too much exuberance from NVDA bulls. We could start with the well-documented moment earlier this summer when Nvidia CEO Jensen Huang autographed a women's undergarment at a technology conference. This is the type of thing typically seen from entertainment industry celebrities rather than technology company CEOs. More recently, the company's latest earnings report had an unofficial watch party at a New York City bar. Which is again an activity that is highly unusual for Wall Street participants and more akin to something sports fans would do during a big game or event. Perhaps these signals could simply be viewed more as symptoms of high interest from retail investors. In the chart below, I'm showing the retail trader active percentage with vertical orange lines indicating earnings reports. It's very common to see spikes in retail trader activity immediately following earnings. However, the last two quarters have shown sustained growth in retail involvement. In the second half of 2023, retail trader activity percentage was rarely above 6% during most sessions. Since the beginning of March, retail trader percentage has been below 6% during just 4 sessions. In fact, retail trader activity percentage has not been below 10% since late-July. The point should be clear here; retail is in this trade in a big way. In my view, this is a signal that we're near the cycle top if we haven't made it already. I make no assertions that I'm an expert technical analyst by any stretch of the imagination. When I use TA, I try to keep it simple and utilize things like moving averages over long periods of time. I see several potential warnings signs with this daily NVDA chart: First, NVDA hasn't sniffed its 200-day simple moving average in well over a year and half. My base case is that we see NVDA test that level in the weeks ahead. That MA is shown via the green line in the chart above, and it is 17% below current levels as of article submission. This would imply a 34% decline in NVDL if the stock does indeed revisit that moving average. Perhaps coinciding with a possible 200-day MA test would be the early August low of $90 per share. It's certainly within the realm of possibility that the stock could look to that level for support. But there are additional issues. For instance, we now have two lower highs from pullbacks going back to mid-June, indicating a more significant topping pattern. And there is an enormous gap that has still not been filled from May 2023. There is no rule that says all gaps must be filled. But given how overvalued this stock is, I don't think it can be completely ruled out. Nvidia is currently trading at 21x forward sales. This is a 650% premium to the info tech sector median. In fairness, these P/S ratios are a testament to how stellar the company's gross margin has been in the race to scale AI compute. But even switching to P/E multiples, NVDA can't reasonably be called "cheap." At 38x forward earnings, NVDA trades at a 40% premium to the info tech sector median and is the second most overvalued company in the Magnificent Seven based on that metric after only Tesla (TSLA). There are additional concerns pertaining to NVDA as well. We've seen insiders selling stock at an accelerated rate recently. Since mid-June, Jensen Huang has sold 4.8 million shares of NVDA for $580 million. To be clear, Huang is still a major shareholder in the company and these sales represent a minimal portion of his total holdings. Year to date, insiders have sold $1.5 billion in company stock, with the majority of those sales coming in the past 3 months. We've also learned of a subpoena as the United States DOJ looks into possible antitrust law violations by the company. Please review the risks of leveraged ETFs as explained in the linked SEC bulletin before placing any trades in NVDL. These risks include extreme volatility, time and volatility decay, and the real possibility of losing all of your investment. The normal risks can be exaggerated with single-stock ETFs such as NVDL. Only experienced traders who fully understand the risks should trade them. This article further explains the important risks of such ETFs with examples. NVDA has been on a truly breathtaking rise over the last two years. Given that rise, it is perhaps unsurprising that retail has found its way to the stock to such a large degree. But there is a fairly obvious warning sign there. Readers are no doubt familiar with the concept of fading rallies when the cabbies start giving stock tips. Perhaps in this era, that sell signal comes from the Uber (UBER) driver. Or maybe earnings report watch parties work just as well. Beyond sentiment, I see red flags on the chart. I see red flags from insider sales. I see a red flag from the DOJ. And I see a red flag from a 2x leveraged ETF that hasn't yet punished people for holding shares since 2022. Things that have a beginning have an end. I'm certainly not saying to short NVDA or to even sell NVDA for that matter. But I do suspect that the company's growth has been priced in for now, and NVDL is unlikely to continue on this meteoric rise. NVDL is a sell.