NVIDIA: Beyond AI - A Comprehensive Look at Q2 Performance and Future Prospects

Nvidia Review: Sequential Revenue Decline In The US (Rating Downgrade) (NASDAQ:NVDA)

Despite the AI revolution, Nvidia's current market cap seems to overprice future growth, prompting my now strong sell opinion. NVIDIA Corporation's (NASDAQ:NVDA) Q2 earnings report on the surface showed the tech company firing on all cylinders, with revenues and EPS surpassing both top and bottom-line expectations, respectively. The chip giant's non-GAAP EPS of $0.68 beat estimates by $0.04, while revenue reached $30.04 billion, above expectations by $1.31 billion. Nvidia's Data Center revenue hit a record $26.3 billion, up 154% year over year. The company also issued guidance for FY Q3 2025 with revenue expected to be $32.5 billion, which is slightly above the consensus of $31.75 billion. Despite what appear to be strong figures, Nvidia's shares fell after hours Wednesday, reflecting what, I think, is the market sniffing outgrowing concerns that have been simmering. For those who haven't followed my coverage of Nvidia, I have become increasingly skeptical about the chipmaker since I started writing about them last fall. My concerns (first pointed out in November) on competition seem to be coming true. The Blackwell chip delay seems to be the catalyst that caused the market share disruption I have been warning about all along. I think the market is adjusting for potential challenges that Nvidia is facing under the hood, such as delayed plans to ship billions of dollars worth of Blackwell chips and concerns about revenue peaking in the United States (as noted in their 10-Q and on the earnings call). While Nvidia is still confident about the Q4 delivery of Blackwell chips, the market's skepticism is definitely higher than it was pre-earnings. There's no doubt in my mind that Nvidia will figure out these chips and scale production. The question is if these chips will ship in the volume they estimate on the Q4 timeline they plan. What's interesting about the admissions from the earnings call is that none of the concerns about a "three-month delay" like I covered before earnings seemed to have been dispelled. I (from my research) have said that Blackwell chips would ship late this Fiscal year. Management confirmed this on the call. But management was vague. Without Blackwell, Nvidia will likely find it tougher to compete against Advanced Micro Devices, Inc. (AMD) given their MI 300X chips are now ramping production in key use cases. Nvidia is heavily dependent on Blackwell's success going forward. I think investors need to weigh the risks and consider the impact of a prolonged delay. With this, Hopper sales (the immediate revenue solution to this delay) actually saw their sales peak in the United States (where the plurality of global Data Centers are located) sequentially 2 quarters ago. While overall Hopper sales increased, the sequential decline in revenue in this key market (I think) runs completely contrary to the Nvidia narrative that demand for Hopper is strong where the big customers are. Most of their large Hyperscaler customers (Big tech) are in the United States. They (on the whole) are buying less. Despite my bearish stance on Nvidia heading into earnings, I actually anticipated slightly stronger Hopper performance this past quarter, given their (I'll admit) strong momentum in AI and data center. I did not expect to see the sequential drop-off in Hopper sales just yet. With this, while some saw the quarter as a beat, I think we actually have some structural issues underneath. I am downgrading my view on Nvidia from a sell to a strong sell. Last month, I wrote pre-earnings coverage that highlighted concerns I had about Blackwell chip delays and cautioned investors about potential revenue disruptions stemming from delays in Nvidia's upcoming Blackwell chips. With the release of their Q2 results, my concerns were partially reflected in commentary on the call. Despite management's admission that the Blackwell chip design had to be reengineered (the foundation of why the chips were delayed), management provided limited clarity on the precise length of the delays, which I think left many investors with lingering concerns. Management talks about billions of dollars of revenue from Blackwell chips in Q4 FY 2025, but did not specify a number. Is this $3 billion? $8 billion? This chip is the future of Nvidia in many ways. The lack of clarity was frustrating. With this, they have deflected concerns by talking about Hopper sales. I will go into this later, but I am not sure about the durability of this demand. As with before earnings, I continue to be a contrarian to the market with my coverage. I know I am severely diverging from the bullish consensus on Wall Street. I think the little details add up here. The point of my follow-up coverage is to show why I am now more bearish and to give investors another viewpoint. As with most quarters for Nvidia over the last two years, one of the key highlights in their Q2 results was the record Data Center revenue of $26.3 billion, which grew by 16% (on the whole) from the previous quarter and 154% from the prior year. We have to dive into these details to get a better view of these sales, however. In their most recent 10-Q filing, it's now clear that sales in North America have actually begun to decline, which could be attributed to potential market saturation, waiting for the Blackwell series, and increased competition within AI and data centers (see my research on AMD). This is compared sequentially to FY 25 Q1. When asked about this on the call, Nvidia's CFO responded that where a customer's revenue is booked may not be where the customer ends up using the product (as a deflection for why US revenue declined sequentially): That's not necessarily where the product will eventually be, and where it may even travel to the end-customer. These are just moving to our OEMs, our ODMs, and our system integrators for the most part across our product portfolio. What's notable about this quote, however, is that she is referring to her Original Equipment Manufacturers (OEMs) & ODMs (Original Design Manufacturers). Think companies like Super Micro Computer, Inc. (SMCI) and Dell Technologies Inc. (DELL). This does not appear to be the big tech companies we think of that have been some of the biggest buyers such as Microsoft Corporation (MSFT) or Meta Platforms, Inc. (META). So, a sequential revenue decline can only be partially explained by OEMs and ODMs shifting purchases to non-US locations. I, personally, think ODM demand is less sustainable than direct-to-user demand like that of Big Tech. ODMs are technically resellers of Nvidia chips. If they see demand lighten up, they will cut orders to keep inventory lean. Big tech has already indicated (from their earnings calls) that they will overbuild for now. Note as well that Nvidia's CFO on the call provided no statement to analysts on whether this sequential drop would be just a blip or part of a new trend. My bearish belief (hence this article) is that this is not just a blip. She also implies this sequential sales drop is due to GPU/datacenter sales (vs. automotive or other category sales) because she said it was companies who build data center servers that were shifting their invoices. Besides my concerns about Hopper sales, the Blackwell chip delays were touched on during the earnings call, and none of the executives were specific about the nature of the defect. Instead, CFO Kress & Huang stated: 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. -Kress The change to the mask is complete. There were no functional changes necessary. And so we're sampling functional samples of Blackwell -- Grace Blackwell in a variety of system configurations as we speak -Huang. It's bullish for Nvidia that they are figuring out how to scale the Blackwell chip. What is not bullish (in my opinion) is the level of vagueness on the magnitude of sales we will see from Blackwell this year. Despite the selloff following earnings, Nvidia shares continue to command a valuation far exceeding the sector median. Nvidia's forward P/E ratio is approximately 41.33, which represents a 73.91% premium over the sector median forward P/E of 23.77. While growth overall is strong, I really think their extended P/E multiple over the sector median really raises concerns now about the sustainability of where the stock is. This is especially true in an environment where growth expectations are being scrutinized by analysts and investors more rigorously. I, personally, found the analyst questions on this earnings call to be far more probing and specific than previous earnings calls. Given Nvidia's key performance in the US, I think there is a strong case to be made for a more conservative approach to valuing their shares. A more realistic stance, I think, would advocate for Nvidia trading at just a 25% premium to the sector median forward P/E. If Nvidia's forward P/E were adjusted to a 25% premium, this would imply a forward P/E of approximately 29.72, compared to the current forward P/E of 41.33. Specifically, if the market followed through and re-rated shares for this adjustment, it would result in a downside of approximately 28.10% for shares on top of the post-earnings drop. I truly believe the market may have overpriced Nvidia's growth prospects relative to the sector. During the earnings call, Huang emphasized that over $1 trillion worth of AI server infrastructure needs to be upgraded. He noted that as servers move away from CPUs to accelerated computing powered by GPUs, there's a growing demand for more efficient and powerful computing solutions. He said on the call: ...the world is moving from general purpose computing to accelerated computing. And the world builds about $1 trillion dollars' worth of data centers -- $1 trillion dollars' worth of data centers in a few years will be all accelerated computing. In the past, no GPUs are in data centers, just CPUs. In the future, every single data center will have GPUs. And the reason for that is very clear is because we need to accelerate workloads so that we can continue to be sustainable, continue to drive down the cost of computing so that when we do more computing our -- we don't experience computing inflation. We are at the beginning of our journey to modernize a $1 trillion dollars' worth of data centers from general-purpose computing to accelerated computing. If the opportunity is as vast as suggested, it begs the question of why the United States, which houses more data centers within its borders than any other country, had revenue that dropped, let alone slowed notably last quarter. I get that some ODM revenue might have been booked overseas, but in the US it just doesn't make sense that revenue would even come close to peaking if the opportunity was so strong. GPU demand is seasonality agnostic (so this is not a seasonal fluctuation) and according to Huang, we are at the beginning of a computer revolution. With this, while the potential $1 trillion opportunity is undoubtedly massive (and maybe I could be wrong it could be this big), the company's current market capitalization is roughly three times the size of this entire opportunity. Usually, the total addressable market, or TAM, of your market is bigger than the net present value of future profits from the opportunity (the market cap). Don't get me wrong: the AI revolution is here and Nvidia will play a big part in it. I just think that Nvidia's shares appear overvalued despite the opportunities in AI. The discrepancy between the potential market and Nvidia's current valuation suggests (to me) that the stock price already prices at an unrealistic level of success, leaving little room for error or market disruptions. While beating estimates on the top and bottom lines, Nvidia Corporation's valuation remains a concern to me, especially given the sequential decline in their United States revenue. With no Blackwell sales expected for this current quarter, the near-term outlook appears bearish (and this could compound into the valuation). The combination of these factors - potentially peaking US demand, and delayed new product contributions - reinforces my view that Nvidia's shares are overvalued, prompting my strong sell belief.

Nvidia: Not Just An AI Play (NASDAQ:NVDA)

I am initiating my coverage of NVIDIA Corporation (NASDAQ:NVDA) in this article. More specifically, I dissect the company's latest earnings report and argue why it's not just the AI revolution that's going to drive the company's future growth. It was yet another blowout quarter from NVDA, in my opinion. Q2 Revenues came in at $30.04 billion, which represents a staggering growth of 122.4% y/y, beating analyst estimates by $1.29 billion. Data center revenues, a closely watched metric, came in at $26.3 billion, up 155% y/y and 16.4% q/q. Gaming revenues bounced back, coming in at $2.88 billion, up 15.7% y/y and 8.8% q/q. Non-GAAP diluted EPS came in at $0.68, which translates to an even more astounding y/y growth of 152%, and also represents a q/q growth of 11%. The EPS print also beat analyst estimates by $0.04. Q2 gross margins came in at 75.1%, up 5% y/y but down 3.3% q/q. Operating expenses saw a sharp increase, coming in at $3.9 billion, up 48% y/y and 12% q/q, primarily due to higher compensation-related costs. The company generated $14.5 billion in cash flow from operations. Management's Q3 guidance also beat analyst estimates. Q3 revenues are expected to come in at $32.5 billion (+/- 2%), much higher than analysts estimates of $31.7 billion. Q3 gross margins are expected to be between 74.4% and 75% (+/- 0.5%), with full-year gross margins expected to be in the mid-70% range. The Q3 operating expenses are expected to be $3 billion, with full-year operating expenses expected to grow in the mid-to-upper 40% range, primarily due to the company's efforts in developing the "next generation of products." The company is expected to generate other income of approximately $350 million, and non-GAAP tax rates are expected to be 17% (+/- 1%). Finally, the company also announced that its board had authorized $50 billion in share buybacks. NVDA's second quarter results, especially the 155% y/y growth seen in the company's data center division, confirmed what its fellow Sensational Six members (Amazon, Apple, Meta Platforms, Alphabet, and Microsoft) highlighted during their respective earnings calls: the rate of AI spending is not slowing down any soon. The cloud service providers accounted for approximately 45% of NVDA's data center revenues. Furthermore, even as many of them wait for NVDA's next-generation architecture, Blackwell, which the company plans to ship only by Q4, for now, its Hopper architecture is more than enough for these CSPs for now. None of them likely want to put off their AI spend until then. And even after the company starts to ship Blackwell, NVDA still expects its Hopper shipments to increase in the second half of FY25, given that the demand for the former is well above what the company can currently produce. In addition to the usual suspects, one area, in particular, which showed considerable growth during the current quarter was Sovereign AI. As the AI revolution moves up a gear, data privacy and security could be paramount, not just for organizations but for entire nations. It's a development, which NVDA management is already starting to observe. During the earnings call, management did provide an example of how Japan's National Institute of Advanced Industrial Science & Technology has partnered with the company to build its AI-driven supercomputer. The company now expects sovereign AI to already hit low-double-digit billions this year. In my opinion, the potential for this segment is massive. As the saying goes, "data is the new oil," and NVDA's management is, unsurprisingly, seeing that countries are considering their data to be a "national resource," and are therefore, building their own infrastructure to have their "own digital intelligence." The US Government has also started the process of geofencing the country's AI tech for strategic purposes and to ensure AI safety. For instance, the US Department of Commerce's AI Safety Institute has signed agreements with the likes of Anthropic and OpenAI to collaborate on "AI safety research, testing & evaluation." According to these agreements, the AI Safety Institute will have first access to the major new AI models from these companies before their public release. Based on this development, the idea of nations funding their own AI startups in the future, based on their proprietary data, is not farfetched at all. The enterprise AI wave, as NVDA management calls it, has already started. But Sovereign AI is one area that has the potential to really accelerate the AI revolution. And based on current evidence, NVDA, and its Blackwell architecture, stand to be a major beneficiary of this trend. Nvidia's evolution from a gaming company to a data center company may have been complete thanks to the AI revolution, but it does not mean that the company's other segments, including gaming, should be ignored. One division of the company, which is slowly gaining pace and momentum, is automotive. The company's automotive and robotics segment, during the quarter, generated $346 million in revenues, an impressive growth of 37% y/y. It also translates to a q/q growth of 5%. The growth was primarily driven by ramps in self-driving platforms and a surge in demand for AI cockpit solutions. As autonomous vehicles become the next battleground for major automakers, with many of them moving towards building Robotaxis, Nvidia's automotive business could be a major beneficiary in the coming years. During the earnings call, management touched upon this by confirming that automakers developing AV technology are all using Nvidia in their data centers. Management expects the automotive segment to be a multi-billion dollar business and attributes this to the potential surge in demand for computing power from next-generation AV models. They are certainly not wrong on that front. In my last article on Uber Technologies, Inc. (UBER), I wrote at length about the future of AVs. I foresee the growth in this market as a major catalyst for UBER and explain why its AV partnership with BYD Company Limited (OTCPK:BYDDF) is only the start. Furthermore, as I mentioned in the same article, according to Fortune Business Insights, the global AV market size is expected to surge to $13.6 trillion by 2030 at a CAGE of 32.3%. And NVDA could be a major beneficiary of this revolution, in my opinion. NVDA's advanced driver-assistance platform, DRIVE's latest version, Thor, which was launched in March of this year, has been widely adopted by major automakers who have AV ambitions. Major Chinese automakers such as BYD, Hyper, XPeng, Nuro, Waabi, and WeRide have all adopted Thor, which integrates the company's Blackwell Architecture. Li Auto and ZEEKR have also announced that their AV roadmap will also be built on the Thor platform. The AV revolution in China is well ahead of the US, and companies such as BYD are continuing to spend billions towards autonomous driving. The Chinese AV market, according to Research And Markets, is expected to hit $31.6 billion by 2030 at a CAGR of 21.7%. And with the Chinese players also harboring European ambitions, the growth potential for AVs, despite the tariffs, is huge in my opinion. Today, Nvidia may be synonymous with AI, but that does not mean that investors should discount its other divisions, especially automotive, which is primed to be a major catalyst for the company in the future. Source: Company's Q2FY25 Press Release, LSEG Data (formerly Refinitiv), Seeking Alpha and Author's Calculations. The company now expects third quarter revenues to come in at $32.5 billion (+/- 2%). The high-end of this guidance would put Q3 revenues at $33.1 billion, a y/y growth of 83.1% and q/q growth of 10.2%, which would be a deceleration from the previous two quarters (15.3% q/q in Q2 and 17.8% q/q). Q3 gross margins are projected to come in at 75% (+/- 50 bps). I have assumed Q3 gross margins of 75%, which results in Q3 gross profits of $24.8 billion. Non-GAAP operating expenses are projected to come in at $3 billion, and other income is expected to be about $350 million. I have also assumed interest income to be the same as Q2, which is £383 million. The company assumes non-GAAP tax rates for the third quarter to be 17% (+/- 1%). I have assumed Q3 tax rates to be 16%. Taken things together, we get the projected net income (after taxes) for the third quarter to be $18.93 billion. The weighted average shares used to calculate diluted EPS stands at 24.85 billion, which results in projected diluted EPS of $0.76 for the third quarter. Since the AI revolution began back in 2023, on average, NVDA has registered q/q growth of 29.43%. If you remove the second quarter of FY23, where NVDA registered a q/q growth of 87.9%, an outlier in my opinion, the average drops to 19.7%. This is a more reasonable estimate for Q4 revenue growth, given that the company expects to ramp up production of Blackwell during the quarter and given that demand for Hopper continues to be strong. At a q/q estimated growth of 19.7%, the company's Q4 revenues are projected to come in at $39.6 billion. I have assumed Q4 gross margins to be 74%, a slight decline from Q3 due to the production ramp of Blackwell. This would result in Q4 profits of $29.3 billion. The company expects full-year operating expenses to grow in the mid-to-upper 40% range. I have assumed the growth rate to be at 45%, which would result in FY24 operating expenses, on a non-GAAP basis, of $11.35 billion. Including Q3 projections, total operating expenses at the end of the third quarter are projected to be $8.3 billion, which implies that Q4 non-operating expenses are projected to come in at $3.05 billion. I have kept my assumptions for interest and other income for Q4 to be the same as Q3, which are $383 million & 350 million respectively. I have assumed an effective tax rate for Q4 of 16%. Taken things together, this results in Q4 net income (after taxes) of $22.7 billion. The weighted average shares used to calculate diluted EPS stands at 24.85 billion, which results in projected diluted EPS of $0.91 for the fourth quarter, which would subsequently result in FY24 diluted EPS of $2.96. The company, according to LSEG data (formerly Refinitiv) currently trades at a forward P/E of 34.2, similar to its 10-year historical median multiple. Given the potential growth, not just in its data center business but also in its other divisions such as automotive, I have assumed a higher multiple of 35.6, which is the company's 2-year historical median multiple. The company, according to Seeking Alpha, currently trades at a forward PEG ratio of 1.15, significantly below its 5-year average of 2.09 and even below the sector median of 1.93. However, I do believe that its current ratio of 1.15 is justifiable, given the multitude of catalysts for its earnings growth. As such, I have assumed this multiple for my calculations, and with a forward P/E of 35.6, the projected earnings growth comes to 30.96%. At this earnings growth, the company's FY26 EPS comes in at $3.88. A forward P/E multiple of 35.6 and a projected EPS of $3.88 results in a price target of $138, which suggests an upside of approximately 16% from current levels. The main risk factor, in my opinion, is the slowing sequential growth in revenues and gross margins. While there were valid reasons, such as inventory provisions for low-yielding Blackwell material, for the decline this quarter, whether the company can reverse its growth trajectory is an area that investors should watch out for. Furthermore, the rise in competition is also a headwind to growth, and at what rate competitors like AMD are grabbing market share will also be key for NVDA's future growth. Nvidia Corporation had yet another impressive quarter, with guidance also nearly beating expectations. The growth expectations for this company have been so stratospheric in recent times, that the magnitude of the guidance beat is what has been driving the stock price. As such, the stock tanked after its earnings release since the magnitude didn't match investors' already lofty expectations. If one digs a little deeper, however, then one would see a company that's firing on all cylinders. The company's data center business confirmed what the likes of Microsoft and Alphabet talked about, which is that AI spend is here to stay and that Nvidia Corporation stands to be a major beneficiary of this development. At the same time, the rise of Sovereign AI is an area that investors really need to pay attention to, given that it has the potential to be a substantial growth driver in the coming months. Lastly, the AI revolution should not distract investors from the company's other divisions, especially automotive. As automakers pivot towards building autonomous vehicles, the automotive division of NVDA is set for years, if not decades, of growth. When one really thinks about it, there are two revolutions taking place simultaneously: the much talked about AI revolution and the quiet AV revolution. And NVDA, in my opinion, is one of those rare companies that stand to benefit from both. As such, one can't ignore this semiconductor giant for too long.

Nvidia Q2: Don't Forget Software (NASDAQ:NVDA)

Looking for a portfolio of ideas like this one? Members of Envision Early Retirement get exclusive access to our subscriber-only portfolios. Learn More " My last article on NVIDIA Corporation (NASDAQ:NVDA) was titled "Nvidia Q2 Preview: Blackwell And Liquid Cooling Could Form Killer Combination." As the title suggests, it was published shortly before the release of the Q2 earnings report and was meant to be my preview of the report. As you can guess from the title, the preview emphasized the hardware side of the business, in particular, the potential of its Blackwell chips. Quote: Nvidia's key partner, Super Micro Computer, recently reported the delivery of direct liquid cooling (DLC) systems for NVDA chips. NVDA's Blackwell AI chips, when combined with DLC, could fundamentally shift the cost structure for AI developers and end users. I expect to see this catalyst reflected in NVDA's upcoming Q2 earnings report. The potential energy saving and performance boost are so significant and thus will materially widen NVDA's long-term competitive moat in my view. Since that writing, the company released 2nd quarter FY25 financial results on August 28, and I thought it would be helpful to review the earnings report. More than a dozen SA authors have written review articles on the stock in the past few days. Like my preview article, most of these articles focused on its hardware. Therefore, my goal in this earning review is to concentrate on the development of its software ecosystem. A key lesson I learned from other established companies (such as Microsoft, Adobe, Oracle, etc.) is that the software ecosystem can form a more durable moat than hardware. Judging by the results it reported in Q2, I think the NVDA is progressing well in this direction too. Before I dig in, let me first recap the Q2 financial statements to better prime the remainder of the discussion. The company reported another quarter of robust growth (see the next chart below) with results beating consensus expectations. As seen, its revenue grew 122% YOY and surpassed $30B for Q2. The growth was led by the Data center segment, with YOY growth of 154% thanks to the strong demand for its AI chips. Looking ahead, the anticipation for its next-gen Blackwell chips is incredible, as the CEO and CFO commented in the release. Against that background, let's move on to the software side now. With its hardware dominance well-known and a key anchor for most bullish theses, I will argue that its software should be a key factor too - but often underappreciated by many bulls. As just mentioned, a key lesson I learned from the IT industry is that the software ecosystem can form a more durable moat than hardware. Microsoft's history is an excellent example. Its enduring success is largely attributed to its robust software ecosystem, in my view. This ecosystem, built around its Windows operating system and various productivity tools, has proven to be a far more durable moat than any of its hardware-based initiatives and most of the technology companies that are hard-oriented. The underlying reasons are also elementary and powerful. First, software products enjoy superb scalability (or the so-called Network Effects). Microsoft Corporation (MSFT) selling Windows to 100 million people does not cost much more than selling it to 10 million people. Yet, the more people use Windows, the more software developers create applications for it. This creates a virtuous cycle where the platform becomes more valuable over time. As a direct result of such scaling and network effects, a well-established software ecosystem is highly sticky, and the switching costs can be very high. Take Microsoft as an example again. Once users and businesses are deeply invested in the Microsoft ecosystem, it becomes difficult and costly to switch to competing platforms. The interconnectedness of Microsoft's products, coupled with the time and resources required to migrate data and retrain employees, creates a significant barrier to entry. Back to NVDA, judging by its recent developments, I am impressed by its capability to help customers quickly scale and deploy their services, often AI-related and highly synergistic to its hardware chips. I anticipate such recurring software services to be the next growth bright spot for NVDA in the years to come. Specifically, NVDA's NIMs (Inference Microservices) are a highlight on this front. These NIM software containers offer a range of crucial capabilities (see the next chart below), and they can tremendously shorten NVDA's enterprise customers' development time and speed up their AI application deployment. Furthermore, NVDA has already developed a good pricing model as quoted below: Nvidia's pricing ladder: A lower barrier to the adoption and deployment of AI inferencing has upsides for both software licensing and hardware sales. On the software side of things, the AI Enterprise license necessary to deploy NIMs in production will set you back $4,500 per GPU per year, or $1 per GPU per hour. The features and pricing model have made the NIM highly attractive to many well-known names. More than 200 technology partners are integrating NIMs for their domain-specific AI applications. These patterns are full of sector leaders spanning manufacturing, healthcare, financial services, retail, customer service, etc., as illustrated by the examples quoted below: Foxconn - the world's largest electronics manufacturer - is using NIM in the development of domain-specific LLMs embedded into a variety of internal systems and processes in its AI factories for smart manufacturing, smart cities and smart electric vehicles. Lowe's - a FORTUNE® 50 home improvement company - is using generative AI for a variety of use cases. For example, the retailer is leveraging NVIDIA NIM inference microservices to elevate experiences for associates and customers. Siemens - a global technology company focused on industry, infrastructure, transport and healthcare - is integrating its operational technology with NIM microservices for shop floor AI workloads. It is also building an on-premises version of its Industrial Copilot for Machine Operators using NIM. In terms of downside risks, Nvidia faces several risks that are common to its AI-related peers. These risks include high valuation multiples (42x FWD P/E as of this writing), intense competition from other AI chip manufacturers (such as AMD and Intel), sensitivity to global supply chain disruptions due to the specialized components required for advanced chips, sensitivity to geopolitical risks (e.g., due to the trade-tension between the U.S. and China), etc. These risks have been thoroughly discussed in many other SA articles (and my own past articles too). So I won't further elaborate on any of them here anymore. Instead, I will point out a risk that is less often mentioned: insider selling. The chart shows insider selling activity on NVDA stock over the past three months. As seen, the recent insider transactions were completely dominated by selling. To wit, a total of 397 selling transactions were disclosed by the insiders in the past 3 months and these sales resulted in a total transaction value of over $1 billion. Notably, the CEO, Jen-Hsun Huang, has reported a series of selling transactions recently. As seen, his recent sales in August 2024 were largely made around an average price of $105 per share. Such significant insider selling activity could add to the selling pressure, create resistance to the upside movement of the stock prices (especially in the near term), and signal a valuation concern from the insiders. All told, NVDA is certainly not suitable for all investors. I see large uncertainties on both the upside and downside and expect large price volatilities to persist. However, for investors with the risk tolerance to stomach these volatilities, I see an attractive reward/risk curve. Besides the dominance of AI chips (the focus of my last article), I see nonlinear growth potential on the software front, too. Notably, Nvidia Corporation's NIMs are progressing quickly, judging by recent financials and customer acquisition. I anticipate the synergies between its hardware and software to form an even wider and more formidable moat in the long term.