Semiconductor Industry Faces Mixed Fortunes: Intel Struggles While Nvidia and Supermicro Show Promise

Intel Stock: Will The Collapse In Confidence Persist (NASDAQ:INTC)

With all of the above and a pressured cash flow, I remain very concerned about Intel's performance over the coming years. Investing in semiconductor exposure has been an exciting journey for most, as exemplified by the development of the iShares Semiconductor ETF (SOXX) which has returned more than 600% over the past decade or more than 220% in just the past five years. The five-year return of SOXX exceeds the ten-year return of the S&P 500, which comes in at 170%. However, once we dive into the individual companies, some exceed, and others disappoint. Such is the nature for there to be an average, and today I'm looking at Intel Corporation (NASDAQ:INTC) which belongs to the list of companies who don't perform on par with its peers. Over the last decade, Intel has lost 42% of its value, with 57% lost if we zoom in on the past five years. I have to admit, I've been a bit tempted by the prospect of investing in a beaten down Intel over the past years, as I've increased my semiconductor exposure. After all, it is one of the juggernauts of modern-day semiconductor manufacturing, or perhaps, used to be. In the summer of 2021 I published an article called "You Can't Own Enough Semiconductor: Here Is Why And What", where I, to no surprise, argued in favour of going heavy into the sector, which makes Intel a natural prospect given its status as a key player in the industry. I never did as I was concerned with the development, and having digested the latest quarterly report, I'm now firmly in the camp of those saying it's on the decline in terms of its position in the industry and therefore also its ability to be a long-term compounder. I often pick up beaten down blue chips as stock-related news tend to be over projected and immediately impact the stock more than its real impact over the mid- to long-term, but in this article, I'll highlight why I don't believe that to be the case with Intel. If I were a betting man, I would anticipate the stock to have fallen further in a year's time rather than seeing it appreciate, but of course, it can go both ways. That might be crazy talk to some, but don't forget that just because it's down 50%, doesn't mean it can't fall another 50%. Ultimately, I conclude the market has experienced a collapse in confidence with both Intel and its leadership team, led by Patrick Gelsinger as CEO. A collapse isn't something that speaks in favour of an appreciating stock, and I conclude that Intel has had a lot of chances over the past years to change its trajectory, which it hasn't and now, experiencing pressured cash flows and reduced R&D spending outlook will mean it can be concluded it has lost the battle with its stronger peers, led by Taiwan Semiconductor Manufacturing (TSM). Intel has been consistent with disappointing shareholders in recent years, and the company is living in a period of free cash flow hemorrhaging that leaves investors even more nervous. There is simply too much risk in the case at this point for it to be attractive, and there are healthier alternatives in the market. If you open the Q2-2024 presentation laying out the quarterly results, you are met by an executive summary where the headline theme is the initiation of a $10 billion cost reduction plan for 2025. This is supplemented by a plan of reducing headcount substantially, upwards of 15,000 or more out of a total close to 130,000. In recent times, investors have gotten used to seeing Intel disappoint, and yet, the company managed to surprise to the downside for this quarter, causing an 18% intraday downwards move. Intel's peers, Nvidia (NVDA), Taiwan Semiconductor (TSM) and others experience explosive topline growth, while Intel experiences the exact opposite. As you will see in a minute, Intel has the R&D budget to be competitive, and its scale warrants it should be fighting for the throne, but it just doesn't and hasn't for quite a few years at this point. This industry was always about three things. First, it's about process knowledge. The only way to reach the next plateau is to master the one before it, so if you want to manufacture a 5nm chip, you need to be able to manufacture a 7nm chip. Process knowledge is king, and process knowledge requires people, machinery and software. Secondly, it's about capital spending. This industry is capital intense like none other. Foundries come at a cost that easily exceeds $10 billion per foundry depending on its size, scale and purpose. As a company, you need a strong operating cash flow that will allow you to continue pouring money into running the machinery and stay at the edge of physics, which is basically what semiconductor manufacturing has become in this century. Thirdly, it's about R&D spending as you need to constantly push the boundaries of what's possible, and this again requires human capital and funds. A $10 billion cost reduction plan might help alleviate cash flow challenges, but when you are already in the bottom of the pack, it doesn't set you up for success long-term. As I tried to suggest, this is a marathon where it's difficult to claw your way back once you are trailing. If you look at the development in revenue and R&D spending over the past five years, then it's very evident that they are moving in the wrong direction. Intel keeps plowing ever more cash into R&D, but to little avail in terms of turning that into top-line expansion. Yes, R&D is a long-term game same as Pharma, and yes, it's not entirely fair to compare to TSM, but please keep in mind that this provides a high-level example of how the development is far from what it should be. You need your continuous R&D spend to turn into a positive top-line, especially when you are growing your R&D spend. With no expansion in the top-line, and an increased capital expenditure, there is only one result - diminishing free cash flows. The picture above tells a clear story of a company in a challenging spot. Even if they did improve the situation, would they be able to allocate that into value-adding R&D that would elevate the business, so far, they have struggled with exactly that. Before we shut down Intel outright in this piece, let me just point out that they have +$11 billion of cash on the balance sheet in combination with $18 billion in short-term investments, so the lights still come on in the morning, but this isn't sustainable in my opinion. However, it doesn't have to be sustainable for much longer either, if they just correct the situation, which is naturally what is being communicated as happening just around the next corner. Any management team believes in their own strategy, and there is of course a pitch to follow. Having seen such blatantly poor execution in recent years, however, it's time to look beyond the pitch. There is, however, one positive story to be told, and that is related to Intel, of course, being a sum of business units, and not all are doing bad. The picture above, provides the helicopter perspective of Intel, and that doesn't look appetizing with declining top-line and drastically reduced gross margins meaning that Intel come out with more or less nothing on the bottom line. In addition, the gross margin is quite a bit lower than what the company guided just recently in the prior quarter. We have to look a bit deeper, however, as some of the business units are actually doing okay. Intel operates in a number of business units, the major ones being the following, which are those that can turn the needle for the company. As can be observed, Intel has a strong PC-related business that is supporting especially the negative development in foundry, as can be seen below. What can also be observed, is that its data center business isn't reporting anywhere near the same development as its peers. Intel's product line-up appears to be lacking in competitiveness, despite the very strong demand for anything data center and AI related. That is concerning, as it's here growth should come from. Here is what Pat Gelsinger, the CEO, had to say during the earnings call in relation to the foundry business. "Our investments in a global footprint of leading-edge capacity continues to weigh on near-term profitability, but long term, they position us to profitably participate in the largest and fastest-growing parts of the semiconductor market. We continue to expect the investments we're driving through this year to put us on a course for meaningful financial traction with operating profits for Intel Foundry troughing in 2024 and then driving to breakeven." This isn't unusual in the sense that as a semiconductor company, you have to go through cycles of investing in capacity upfront in order to harvest returns down the road. Texas Instruments (TXN) is doing the exact same thing at this point in time, but their business is performing better on a relative level. In the quote above, Pat Gelsinger concludes that the business will be much better off in 2025, and while that might be true, it might not as well. I'll come back to this in the next part of the article, as the company hasn't managed to deliver on its previously communicated strategy. David Zinsner, the CFO, elaborated "Foundry operating loss of $2.8 billion was worse sequentially. We expect operating losses to continue at approximately the same rate in Q3 with more than 85% of wafer volumes still coming from pre-EUV nodes with an uncompetitive cost structure and power performance and area deficits reflected in market-based pricing. The continued ramp of our Intel 4 and Intel 3 Ireland facility and elevated R&D and start-up costs to support the rapid progression of our leading-edge technology development will also weigh on profitability." In the world of banking, you say that debt is solid as rock, and assets are soft as butter, meaning that you must be careful in terms of what value you assign to an asset. Similarly, R&D spending is a fact, but future revenue is an unknown, and if you are a business that continuously delivers, then investors will ask few questions, but if the opposite is the case, then that calls for caution. What the two executives state here is not in any way illogical, but it is of course a question of whether we take it as facts or not. Facts are, that the business is struggling and has been for a number of years, and they now need to take a deep cut into their capital investment program, which is the one thing fueling Intel's future. I remain a skeptic at this point. Naturally, the questions came pouring in during the earnings call, with the word foundry being mentioned a total of 43 times. One thing is apparent, however, Pat Gelsinger keeps believing in Intel's outlook, as he picked up shares following the negative market reaction, something he has done sometimes over the past years. It is often considered a strong indicator if a CEO picks up shares, and it might as well be here. However, Pat Gelsinger has a reported net worth somewhere between double and triple digit million dollars, so adding a couple of hundred thousand dollars, is perhaps not a substantial amount, lessening the strength of that indicator. I previously mentioned that Pat Gelsinger and his team expect Intel to see its lowest performance in 2024, with an improved situation in 2025. Let's have a look at the strategy laid out just a couple of years ago. Below is the opening headline of Intel's investor conference 2022. Hindsight is 20/20, and it's easy to come off in a bad position when looking backwards at a failing strategy. At that investor conference, Intel pointed towards the company All those expectations came together in form of an anticipated growth story unfolding, as can be observed below. Evidently, Intel is yet to deliver on the strategy laid out back in 2022. There is a difference between missing the target by an inch, and by a mile, while this situation is closer to the second. The stock has reacted, as could be expected, by continuously trending downwards, as the market has gotten to a point where trust in Intel and its management team is very low. The way the situation has unfolded, a very stressed cash flow can be observed. The cash flow development leaves me very concerned. In fact, the development in operating cash flow and capital expenditure, means that Intel is burning cash at an alarming rate. It has gotten to the point where all these capital expenditures need to pay off. Intel's profitability is struggling, causing a lot of pressure on the cash flow. If you made it so far, we didn't even discuss the suspended dividend. I'm not going to speak at lengths about it, I'll let the graph below speak of the typical consequences. Historically and in general, cutting your dividend isn't something that speaks in favour of future performance. Back in 2023, Intel reduced its dividend by 65% and during the Q2-2024 reporting, it was suspended going forward. This just underlines the cash flow bleed the company is going through in the years to come. Normally I find it meaningful to debate the valuation of a given company. The problem, here, is that so many other themes cloud the validity of any valuation perspective. The management team has struggled with transparency in terms of just how poor the development has been. Just two years ago, the management team committed Intel to its dividend, which has been both severely reduced and now suspended since then. The strategy from 2022 hasn't been executed as communicated. The list goes on with points of concern. As investors, we need to look beyond the pitch from top management and ask ourselves if we still have credibility in their ability to deliver on their vision. At this point, I have lost trust in Intel to deliver. The probability of Intel catching up on the technology curve diminishes for every single quarter of weak execution and even more so, when the company is forced to cut into its R&D and capex programs. Valuation wise, Intel has been beaten so much, that it may very well deliver strong returns at some point down the road. One theme I haven't discussed here, is the fact that Intel's foundry footprint is primarily located in the US, and we all know today, that semiconductor manufacturing has become a national security topic, as such I wouldn't be surprised if there will be continued access to grants that can bolster even a struggling Intel. I believe, however, that uncertainty and risk of continued weak performance, will mean the stock, at best, trades sideways in the coming years. However, cheap Intel might be, I believe it's a stock to best stay away from at this point in time. Even so, for investors who have a keen interest in turnaround situations, this could be the classic situation of trying to catch a falling knife.

Intel: A Potential Value Trap (Rating Downgrade) (NASDAQ:INTC)

Management anticipates achieving profitability and FCF breakeven in FY2025. Intel (NASDAQ:NASDAQ:INTC) delivered very disappointing 2Q FY2024 results, with nearly all financial metrics falling short of market estimates. Despite its goal of catching up with other semi players amid the current AI boom, INTC seems to be moving in the opposite direction. The company not only missed revenue and EPS consensus but also issued a weak outlook for 3Q and beyond. Additionally, the management announced a 15% reduction in workforce, cuts to infrastructure investment spending, and plans to suspend its dividend starting in 4Q FY2024, which justified the post-earnings selloff. Despite the stock's already low valuation before the 2Q earnings, it further dropped over 30%, reflecting significant doubts about the viability of their restructuring roadmap. Although I do not anticipate a clear path to a turnaround until FY2026 or later, I believe the selloff has already priced in this narrative. As I mentioned, this is a long-term turnaround story that could be very bumpy. I downgraded my buy rating to hold on INTC due to my concern about a potential value trap. INTC had a big miss on both revenue and non-GAAP EPS estimates in 2Q, indicating a dollar decline in revenue and flat earnings growth on a YoY basis. In other words, the company returned to "0 growth" territory again. In addition, the company guided an accelerated revenue decline YoY and negative growth in EPS in 3Q, reflecting a further deterioration in its fundamentals. The management admitted that the recovery is slower than expected. Particularly, the company's Client Computing Group ("CCG") revenue (58% of total revenue) was largely impacted by an export license restriction in China. As shown in the chart, the segment growth significantly decelerated to 9% YoY in 2Q FY2024, down from 31% YoY in the previous quarter. The company provides a gloomier 3Q outlook, expecting -8% YoY growth in total revenue, considering the midpoint of the revenue guidance range. The growth rate is expected to be the slowest since 2Q FY2023. Despite a -4% YoY growth in Data Center and Artificial Intelligence ("DCAI"), the company expects sequential modest improvements in 2H FY2025, which may indicate a low to mid-single digit QoQ growth sequentially. This also implies a potential negative growth YoY on CCG segment in 3Q FY2024. The disappointing revenue outlook coincided with the management's decision to initiate a cost cutting plan before 2Q earnings release. In the 2Q earnings call, they announced "reductions across OpEx, CapEx, and cost of sales total well over $10 billion in direct savings in 2025". The chart shows that INTC margins contracted significantly to 38.7% in 2Q FY2024, approaching the lowest level in the company's history. This margin is less than half of NVIDIA's (NVDA) 79% and 27% below Advanced Micro Devices' (AMD) gross margin from the latest quarter. Before the pandemic era, INTC had maintained a 60% gross margin trajectory, highlighting the tremendous competitive disadvantage the company currently faces at the moment. The management attributed the gross margin decline to an unfavorable product mix and more competitive pricing than previously expected due to the accelerated ramp of their AI PC products. Consequently, they have guided for an even lower QoQ gross margin in 3Q FY2024, down to 38%, which would mark a new historical low. The company's heavy reliance on external wafers for ramping AI PC products is expected to continue pressuring gross margins in the coming quarters. Therefore, I believe the company's gross margin will continue below 40% and show no inflection in the near-term. In addition, Moody's downgraded INTC's senior unsecured rating to Baa1 from A3, reflecting an increase in its credit risk. In response, the company is trimming capital investments in 2H FY2024 to approximately $12 billion, representing a 50% YoY decline from FY2023. Looking ahead, they expect a modest YoY increase in capex, with projections ranging from $12 billion to $14 billion in FY2025. Furthermore, INTC plans to suspend its dividend payments starting in 4Q FY2024. The management believe these measures will help reduce leverage and achieve FCF breakeven in FY2025. INTC's long-term goal is to restore its gross margin to the previous level of 60% by 2030, with a target of reaching a 40% operating margin as well. In contrast, NVDA's gross margin is currently close to 80%. In the near term, INTC is pursuing an aggressive cost-cutting plan to achieve breakeven profitability by FY2025. While the company has guided for approximately $20 billion in operating expenses for FY2024, they expect a 12.5% YoY decline to $17.5 billion in FY2025, which is significantly below the previous market consensus of $21 billion. This suggests that the company's margin contraction may not bottom out until the next fiscal year. Given the possibility of its non-GAAP EPS turning negative again in the coming quarter, I believe that using the P/E multiple may not accurately reflect its valuation. Following the recent 30% price drop, INTC is currently trading at 0.77x P/B TTM, which is 65% below its 5-year average. Typically, a multiple below 1x signals a strong buy. However, investors should be cautious of a potential value trap, as continued deterioration in the company's fundamentals could justify the low valuation. As indicated by Seeking Alpha's valuation section, INTC's forward multiples are higher than the TTM basis, suggesting that its financial metrics are expected to worsen. Therefore, being simply bullish on a stock because it appears very cheap is not a reliable investment strategy, because it can be cheap for a decade. Instead, we need to assess whether the company can improve its competitiveness in the industry, which investors may have overlooked. However, based on 2Q earnings, I don't see INTC as being in that position at the moment. Therefore, I'm neutral on the stock and prefer to wait on the sidelines to gauge potential inflection points in the coming quarters, guided by management's comments. In summary. INTC's 2Q FY2024 results reveal that its turnaround plan seems to be heading in the wrong direction, marked by a significant slowdown in CCG revenue and significant margin contractions, leading to a negative earnings trajectory. The management's outlook indicates that conditions may worsen before getting better. To address liquidity issues following a substantial decline in fundamentals and its credit downgrades, the company has aggressively cut operating costs, reduced capital spendings, and even suspend its dividend, underscoring severe operational challenges in short to mid-term. Given these concerns and the potential for a value trap, I have downgraded my rating from a buy to a hold. The stock's 30% decline reflects skepticism about the turnaround narrative, and while long-term opportunities might still exist, it's better to be cautious for now. When everyone questions the viability of its turnaround story, long-term opportunity may not emerge.

Supermicro Q4: A Buy Opportunity After Post-Earnings Drop (SMCI)

Super Micro (NASDAQ:SMCI), also known as Supermicro, reported earnings on August 6, 2024. Despite showing triple-digit revenue growth, investors were disappointed in the company's profitability. The stock dropped 20% the day after earnings. It is also down 22% from my buy recommendation at $806.40 on July 20, 2024, and dropped 35% from my April 5, 2024, buy recommendation. Despite the company's subpar fourth quarter fiscal year ("FY") 2024 results and the resulting stock decline, the thesis for my first buy recommendation on November 15, 2023, remains intact. The stock price is up 113% from that call, compared to the S&P 500's (SPX) 23.28% rise. Supermicro may be one of the biggest beneficiaries of the build-up of Artificial Intelligence ("AI") infrastructure by cloud companies and enterprises interested in building generative AI solutions. Experts continue to forecast significant growth for AI infrastructure over the next decade. A blog on Juniper Networks' (JNPR) website stated: IDC's research predicts that by 2025, the 2000 most prominent companies in the world will allocate over 40% of core IT spending to AI initiatives, driving a double-digit increase in the rate of product and process innovations. However, these new business use cases require new infrastructure, and 2024 will see the beginning of explosive growth in AI infrastructure. Supermicro is already capturing much of that explosive growth. The following chart shows that the company has vastly outgrown the industry average so far in 2024. The company grew revenue in the fourth quarter FY 2024 by 143% over the previous year's comparable quarter to $5.31 billion. Chief Executive Officer ("CEO") Charles Liang said on the fourth quarter's earnings call: To put this in perspective, our Q4 revenue has exceeded the full year revenue of fiscal 2022. Our robust growth is driven by our technology and product leadership in the AI infrastructure market, especially with Generative AI training and inferencing. We have been scaling quickly to secure a large share of AI CSP [Cloud Service Provider] opportunities, deploying some of the largest AI SuperClusters in the world. This company is investing heavily in maintaining its outperformance compared to the industry. Just when you think its competitors are closing in, management keeps coming up with potential game-changing ideas, like its new DCBBS (Datacenter Building Block Solutions) product, which could potentially position Supermicro ahead of competitors like Dell Technologies (DELL) and Hewlett Packard Enterprise (HPE). If Supermicro can maintain its dominance, the company has a high potential upside. Now that the stock price has dropped substantially over the last several months, its valuation has fallen to a more acceptable level. I maintain my buy recommendation on the company. This article will discuss how Supermicro's new DCBBS product will speed up the time it takes for its customers to build new data centers. It will also review the company's fundamentals from its latest earnings report, examine its risks and valuation, and explain why aggressive growth investors can buy the stock at the current price. Supermicro's CEO announced on the company's fourth quarter FY 2024 earnings call that it plans to extend its mission from simply manufacturing servers to designing and building whole data centers. He said on the earnings call: By leveraging our system building block and rack-scale plug-and-play solutions, we help our customers achieve the best time-to-market advantage with new and performance-optimized technologies. Now, we are further expanding this solution to the entire datacenter. With [the] rapid deployment of large-scale AI infrastructure, data centers worldwide are facing power shortages and cooling inefficiency challenges. Building these new AI-ready datacenters traditionally takes a long time, averaging three years for example. Our upcoming Supermicro 4.0 DCBBS, Datacenter Building Block Solutions, will reduce customers' new datacenter build time from about three years to two years. DCBBS will also enable customers to refurbish old and smaller data centers in six months to a year. Liang didn't fully explain what the solution entails. However, he did say, "This new offering will significantly improve datacenters' TTO time-to-online and cost, with full integration of AI compute, server, storage, networking, rack, cabling, DLC liquid cooling facility water tower, end-to-end management software, onsite deployment services, and maintenance." It sounds like Supermicro is evolving into an AI infrastructure construction and maintenance company. Many companies build data centers, including AECOM (ACM), Data Centers Delivered, DPR Construction, and H5 Data Centers. There are also many companies that build servers and networking devices. However, no companies that I can think of builds servers, data centers, and maintains them. If Supermicro succeeds with its DCBBS product, it could further differentiate itself from its competitors. CEO Liang said on the earnings call that the company plans to offer DCBBS this calendar year. Supermicro produced eye-popping FY 2024 annual revenue growth of 110% to 14.94 billion in addition to its impressive fourth quarter revenue growth. However, Supermicro's high revenue growth comes at a cost. The company has found it very challenging to manage its rapid growth, and one way that showed up in the fourth quarter was the rapid decline in the gross margin. The following chart shows Supermicro's fourth quarter FY 2024 GAAP (Generally Accepted Accounting Principles) gross margin was 11.23%, well below the five-year median of 15.83%. This underperformance is one of the prime reasons for the stock's steep drop post-earnings. Management expected some drop off in gross margin and telegraphed that to the market last quarter. Chief Financial Officer ("CFO") David Weigand said in the third quarter FY 2024 earnings call when forecasting fourth-quarter gross margin: "We expect gross margins to be down sequentially as we focus on driving strategic market share gains." Whenever you see management refer to "strategic" customers, it relates to large-scale organizations such as massive cloud providers such as Amazon (AMZN), Microsoft (MSFT), and Alphabet (GOOGL)(GOOG) or companies with massive data centers. Still, management didn't think the gross margin decline would be as large as it was. The company's explanation for the steep gross margin decline was that it miscalculated the demand for the liquid-cooled rack market. Supermicro quickly ramped up its supply chain to meet demand for liquid cooling, resulting in higher costs for speeding up delivery of some parts and a shortage of vital liquid cooling components. This shortage delayed approximately $800 million in revenue to July. The company expects to recognize the delayed revenue in its first quarter FY 2025 report. Management believes margins will improve from its current level back into a range of 14% to 17% as supply chain issues ease and the company has lower manufacturing costs from ramping up its new manufacturing plants in Taiwan and Malaysia. Another thing that should help the company restore margins is that it has been able to raise its Average Selling Prices (ASPs) on its products. CFO Weigand said on the fourth quarter earnings call, "ASPs increased on a year-over-year and quarter-over-quarter basis, driven by the value and complexity of rack-scale Total IT Solutions." If you ever see ASPs decline, there could be issues with profitability coming down the pike. The company's fourth-quarter operating expenses increased 75% year over year to $253 million. The steep rise was due to the company spending heavily to build its workforce and invest in R&D. Lower gross margins helped dent the operating margins. The company's fourth-quarter GAAP operating margin was 6.47%, down 391 basis points from the previous year's June quarter. Non-GAAP operating margin was 7.1% versus 11.3% in the third quarter. CFO Weigand also blamed lower-than-expected operating margins on a "higher mix of hyperscale data center business." The Supermicro CFO said early in his prepared comments on the earnings call, "One CSP [Cloud Service Provider]/large data center customer represented approximately 20% of revenues for fiscal year '24." The two comments imply that one cloud provider has a strong enough negotiating position to garner discounts for Supermicro's products and services, potentially lowering margins. The negative impact of the lower gross margin and higher operating expenses fell to the bottom line. Fourth quarter GAAP diluted earnings per share ("EPS") of $5.51 were below company guidance and missed analysts' estimates by $2.28. Non-GAAP diluted EPS of $6.25 was below company guidance and missed analysts' estimates by $1.89. Supermicro's cash from operations ("CFO") to sales ratio is -16.59%, which means that for every dollar of sales, the company loses $0.17. The fourth quarter FY 2024 CFO loss was $635.28 million. My last article on the company discussed why Supermicro's cash flow has dropped so rapidly; it's the cost of its rapid sales growth. To maintain its rapid revenue growth, the company needs to spend cash to build up inventory in advance of future sales. However, the company invoices its customers upon delivery of its products and services. The customer has thirty days to pay the invoice. The timing mismatch between the company spending money for inventory and receiving cash from the sale creates a CFO and free cash flow ("FCF") loss. Supermicro had a fourth-quarter FY 2024 TTM levered FCF loss of $2.834 billion. Due to the company's cash burn, management raised money through convertible debt and an equity raise in the March quarter. CFO David Weigand said on the company's third quarter FY 2024 earnings call: During the quarter, we raised $1.55 billion from a 0% coupon five-year convertible bond offering due in 2029, net of underwriting discounts and offering expenses. We also raised approximately $1.73 billion in net proceeds from the sale of 2 million shares at a price of $875 per share. The proceeds from these transactions will be used to strengthen our working capital, enable continued investments in R&D and expand global capacity to fulfill strong demand for our leading platforms. Its stock sale dilutes existing investors, and convertible debt potentially dilutes shareholders. This dilution results in decreased EPS. Considering that the company also dilutes shareholders through stock-based compensation, investors may be less willing to award the stock a premium in the future based on valuation methods using EPS. The company reported that its GAAP diluted share count increased sequentially from 61.4 million to 64.2 million. Its non-GAAP share count rose from 62 million in the third quarter to 64.8 million shares in the fourth quarter, primarily due to the stock offering and the convertible bond offering. The company ended its fiscal year in June with $1.67 billion in cash and equivalents and $1.77 billion in long-term debt. It has a net debt of $504.4 million and generated a TTM EBITDA (Earnings Before Interest, Taxes, Depreciation, and Amortization) of 1.3046 billion. Therefore, its net debt-to-EBITDA ratio is 0.39, meaning the company can cover its debt with EBITDA. If this ratio ever exceeds 4.0, the company may be in financial distress. The company has interest coverage of 65.02. Analysts generally consider interest coverage ratios above 3.0 as an indication that a company's operating income can safely cover its interest payments. At the mid-point of Supermicro's first quarter FY 2025 revenue guidance, revenue growth would be 207% over the previous year's comparable quarter. Non-GAAP quarterly earnings growth would be up 148% year over year. If the company meets the midpoint of its full-year FY 2025 revenue guidance, revenue growth would be 87% year over year. These figures for the first quarter and FY 2025 are solid and a potential reason some investors remain interested in the stock. The company also announced a 10-for-1 stock split effective October 1, 2024. This company has multiple risks that could disrupt the thesis for investing in the stock, which may be why it is one of the shorted large-cap companies in the market. A recent Seeking Alpha article states, "Super Micro Computer replaced Tesla (TSLA) as the most shorted large-cap stock in the Americas in July, according to a monthly report issued by data and tech firm Hazeltree." This company is making a massive bet on the growth of the AI infrastructure market. Supermicro has built up a large inventory in anticipation of strong server sales. However, some believe the AI infrastructure market is in a bubble. If that sentiment proves true and Supermicro server sales come in less than anticipated, it could worsen the company's cash flow situation and lead to the company potentially writing down the value of its excess inventory, leading to significant financial losses. The stock price would likely drop significantly. Although Supermicro may be one of the best top-line growth companies in the market, investors should avoid considering it as being in the same boat as NVIDIA (NVDA), a wildly profitable company. In its latest quarter, NVIDIA had a gross margin of 78.35%, an operating margin of 64.92%, and an FCF margin of 49%, which are among the reasons the company enjoys a premium valuation in the market. Supermicro's business has a low margin, and it may be more difficult for the company to avoid having its business commoditized compared to NVIDIA, which is a potential reason the company seeks to differentiate with ideas like DCBBS. Speaking of DCBBS, Supermicro's entry into the data center construction market will bring competition from well-established players in the construction industry. The company also faces the risk of customers failing to adopt DCBBS and the execution risk involved in getting this new product off the ground. As I pointed out earlier in the article, Supermicro has a high customer concentration risk with one large cloud provider, which could hurt the company in two ways. The first way is that when one customer becomes a dominant part of the business, it gains leverage to negotiate discounts and lower ASPs, which could squeeze margins. The danger is that if competitors offer better deals than Supermicro, a race to the bottom could commence, resulting in less profitable business for all AI server manufacturers. The second potential negative outcome is that the customer could eventually switch to a competitor, causing a massive revenue decline. When I last discussed Supermicro, its trailing 12-month ("TTM") price-to-sales (P/S) ratio was 4.0. After the post-fourth quarter FY 2024 stock decline, the P/S ratio is 2.5. Although it's still above many of its competitors' P/S ratios, the company's high revenue growth rate may justify its premium valuation compared to direct competitors. According to Full:Ratio, the Computer Hardware sector has an average P/S ratio of 2.4. The stock price would be $589.24 if the stock traded at the Computer Hardware P/S average. The company's TTM revenue per share is $245.52 as of the end of its June quarter. Multiplying a 2.4 P/S ratio by 245.52 TTM revenue per share equals the stock price of $589.24, down 6.3% from the August 15 closing stock price of $626.69. However, its TTM P/S ratio is backward-looking, while the market is forward-looking. Supermicro's forward price-to-sales ratio is 1.29, well below the information technology ("IT") sector's median of 2.8, suggesting the market underestimates its revenue growth potential compared to the IT sector. Let's look at the company's Price-to-Earnings-to-Growth ("PEG") ratio. Some investors consider a PEG ratio of one fairly valued, a number below one undervalued, and a ratio above one overvalued. Since Supermicro has a PEG ratio of 0.431, the market may undervalue the stock at current prices. Note that earlier in 2024, its PEG ratio was around 3.7, a sign of potential overvaluation. The market may have been willing to award such valuations when it was in an exuberant state of mind earlier in the year. However, as more people examine this company's fundamentals and discover potential issues with profitability, investors shouldn't expect valuations to return to a PEG ratio of 3.7 any time soon. In hindsight, I should have issued a Hold or Sell call in my April 5 and July 20 articles. The stock's valuation was too high, especially considering that, at its best, the company may only achieve a 17% gross margin. Let's look at the forward PEG ratio. The following table from Seeking Alpha shows the company's consensus non-GAAP EPS estimate, growth rates, and forward P/E ratio. Generally, the market considers a stock fairly valued when its forward P/E matches its EPS growth rate. The company would have a forward PEG ratio of one, which some investors consider fairly valued. Since Supermicro's FY 2025 estimated EPS growth rate exceeds its forward P/E, the market may undervalue the stock. It has a forward PEG ratio of 0.34 (18.34 divided by 54.68). If the stock's forward P/E equaled its FY 2025 EPS estimated growth rate of 54.68%, the stock price would be $1868.41, up 197% from its August 15, 2024 closing stock price of $626.69. Still, investors may not award the company a PEG ratio of one or above because of previously discussed risks and issues with the company's margins. So, investors should not expect an immediate stock rebound to new highs. Risk-averse investors should avoid Supermicro. Suppose the market for generative AI solutions takes longer to pay off than many expect, or the company fails to generate enough profits from its AI infrastructure products. In that case, people will be happy they avoided investing in the stock. My last two Supermicro buy recommendations have not panned out. However, I made those recommendations at very steep valuations. Supermicro now sells at a far more reasonable valuation. If you are comfortable with the company's risks and looking for a company with a potentially high upside, consider allocating some of your portfolio to this stock. I maintain my buy rating for Supermicro for aggressive growth investors only.

Nvidia Should Beat Again (NASDAQ:NVDA)

I last wrote about NVIDIA Corporation (NASDAQ:NVDA) (NEOE:NVDA:CA) here on Seeking Alpha in early May 2024, ahead of its fiscal Q1 2025 report. At the time, I argued that despite its recent growth rates and solid profit margins, Nvidia may "face challenges from increasing overall competition in the future, a potential near-term slowdown in demand, and increased competition in the accelerator market." This risk hasn't materialized as the time showed. The stock has continued its wild rally, which looks even stronger after the recent dip. We're just over a week away from the release of the Q2 2025 report on August 28th, but I conclude from some signs that Nvidia is highly likely to beat consensus estimates again. In addition, I don't think management's guidance will be weak. I believe all this should lead to a further rise in the stock price in the medium term -- hence my rating upgrade today. First off, I should like to dedicate a few words to what everyone probably knows already -- the Q1 results. In fiscal Q1 2025 (ending April 30, 2024), Nvidia reported revenue of ~$26 billion, marking a 262% YoY increase and an 18% QoQ rise, exceeding the high end of Nvidia's own guidance by ~$1.5 billion and also Wall Street's expectations by $1.48 billion, according to Seeking Alpha. Non-GAAP earnings were $6.12 per share, up 462% from the same quarter last year and $0.96 higher than the previous quarter -- these figures also beat the consensus estimate of $5.59 per share (that's unadjusted for the 10-to-1 stock split the firm had completed): Nvidia's data center segment -- the key pillar to its phenomenal growth -- achieved sales of $22.6 billion, an increase of 427% over the previous year and accounting for 87% of total revenue. Within this segment, sales of AI computers increased by 478%, driven by increased demand for Nvidia's HGX platform, which powers large-scale language models and other AI applications. Despite supply constraints for its Hopper GPU platform, Nvidia is ramping up production. This is thanks to its partnerships with manufacturers such as Taiwan Semiconductor (TSM). It looks like the company will have more opportunities to meet the considerable demand that has accumulated in the few months since the market realized the importance of LLMs (large language models). The gaming segment was also robust with sales of $2.65 billion, which corresponds to an increase of 18% compared to the previous year. Nvidia's GeForce RTX 40 series graphics processors, which are based on the Ada Lovelace architecture, are also enjoying strong demand. Professional visualization revenue of $427 million rose 45% annually, reflecting a higher sell-in to partners following the normalization of channel inventories. On the other hand, the automotive revenue of $303 million increased just by 3% YoY (8% QoQ), driven by self-driving platforms (while the sequential gain reflected "rising demand for sales of AI cockpit solutions"). Based on what I saw in Q1 numbers, Nvidia continues to be a leader in the field of AI; as this technology becomes more widespread, I believe Nvidia's extensive involvement in this fast-growing industry sets it up for further success. However, what is surprising is that the market still has not fully priced this in -- allow me to explain. What I find strange is that while Wall Street analysts raised their forecasts for the next periods after such strong Q1 figures -- we saw that in the screenshots above -- the extent of these upward revisions seemed too small: Normally, such massive earnings beats as we saw in Q1 lead to massive strong upward revisions, but in Nvidia's case, we don't see enough strength. The less than 1% increase in EPS expectations after the Q1 release makes the final Q2 consensus estimate even more pessimistic than before the Q1 release. The thing is, Nvidia's strong Q1 should theoretically have reduced the level of uncertainty, but for some reason, the market has chosen to ignore this. The management guided for ~$28 billion in revenue for Q2 2025, expecting all of their market segments to show growth. Aiming for about 75% in gross margin (~3% lower QoQ) and with OPEX projected to be $4 billion on a GAAP basis and $2.8 billion on a non-GAAP basis, Nvidia should maintain its superior FCF margin, in my opinion. This should give investors more reason to expect more buybacks. Since fiscal Q1 2023 (6 quarters have already passed), Nvidia's revenue has grown 4.3 times, while the gross profit margin has increased 5.3 times thanks to operating leverage. In other words, CAGR growth rates amounted to 27.5% and 32.1%, respectively. Now the market is forecasting non-GAAP EPS growth of only 4.9% on a QoQ basis -- no matter how we try to explain this with seasonality, I don't think growth will be that modest, especially in light of management's (and Nvidia's major customers') comments that AI is just at the dawn of massive adaptation. According to calculations by New Street Research (proprietary source), CAPEX for AI (expenditure on XPUs in data centers) is expected to increase by 50% in 2025 and by 20% annually between 2025 and 2027. Given the recent problems of potential competitors such as Intel (INTC), I now believe that Nvidia will maintain its leading position in the market for much longer than I previously thought. More specifically, the risk to its supremacy has now become less. So Nvidia will be able to translate the projected growth into its future revenue. But apparently the market is still underestimating this, as the current revenue estimates seem too low to be true. Take a look at the EPS forecasts for the next few years: The market is pricing in virtually no operating leverage and apparently assumes that the company has already peaked on this issue. Logically, however, I think Nvidia should have opportunities to expand its margins even further as it expands its chip production capacity. Therefore, I expect Nvidia stock to trade at a valuation premium to for some time. If we assume that the current FY2026 consensus forecast for EPS of $3.78 is 4% below what is likely (which is roughly the average EPS beat over the past decade), then at a P/E of 40x, I see a "fair value" of $157.3/sh. (12-month target), which is about 26% above the current stock price. So I'll conclude that NVDA deserves a "Buy" rating right before its Q2 release and should be bought on any major dip. There are, of course, plenty of risks to my rating upgrade. First off, I think there is a real chance that we are in an AI bubble. The fact that AI could change our world forever doesn't mean this area is a no-brainer investment field. The rapidly increasing investments and valuations from VCs and others could be due to hype rather than sustainable, long-term changes in demand. If this is the case, there could be a steep market correction, leading to massive losses for Nvidia and its shareholders. Additionally, the ongoing investment cycle in AI and data centers is very supportive of Nvidia, but as I noted in my very first article on NVDA stock, such investment cycles can turn quickly. For example, a recession that results in businesses and governments cutting back on their AI spending -- due to economic pressures, new technological trends, or geopolitical developments -- could imperil Nvidia's future growth prospects. Further, a sudden decline in capital spending from the largest cloud providers and enterprises would result in Nvidia's products facing lower demand, which could also hurt its top-line as well as bottom-line growth. And of course, I can't escape the topic of valuation. The Nvidia stock has already rallied strongly, as high hopes for future growth have driven its price up. So the stock is perhaps already fully priced and there is no margin for error -- any slowdown in growth, a fall in profits or even a slight change in positive market sentiment could have a significant impact on the stock. Despite the possibility of an AI bubble looming, I believe Nvidia's continued leadership in AI and strong financials suggest a positive trajectory in the medium term. I believe Nvidia should beat earnings again (for Q2 2025) because a) the management clearly guided for stronger growth across all segments and b) the current estimates weren't adjusted properly, in my view, after the company released its strong Q1 2025. The stock could be undervalued by around 26% if my premise of maintaining the valuation premium is upheld. So that's why I think that any major dip in the stock should be viewed as a buying opportunity. I thus have upgraded Nvidia Corporation stock to "Buy" about 1 week before its Q2 release.