Called $GOOGL bullish on Sep 11, 2025: $GOOGL rose 41% in 5 months.
View call →When AI makes intelligence cheap, who stands to benefit?
Cite asCryptoQuant has tracked the directional calls on NVDA by @reboundcapital on Substack since Jun 2025: not yet scored on NVDA (as of Nov 19, 2026).
Other assets covered (11)
2 NVDA calls on record, a thin history.
Relative view by default: fewer than 30 NVDA calls or days on record.
Sentiment timeline shows relative sentiment within this analyst's history. A perma-bull showing 8 bullish : 2 bearish in a bear market sets that as their baseline. If you notice any errors, claim @substack:reboundcapital to submit corrections.
Hover the chart to read the thesis of that period · click a day to pin it
Top Rebound Stocks for Nov’26
The markets dropped sharply after the election results and have largely recovered. We may see another round of selling as the Middle East war, the questions over Nvidia’s CEO, and the debate over AI capex in 2027 play out. This may give us another sharp drawdown in the near term. This is how every drawdown will be talked about if the market falls. First, the stock will fall. Then people will come up with a litany of reasons to explain the drop, and the online discourse will become mostly negative. One of these explanations will become the dominant narrative, and the derating can continue for weeks to many years. It may be warranted ( the market is usually right ), but sometimes the market gets it wrong and gives us an opportunity. Buying into these sharp drawdowns is how the Rebound Portfolio has beaten the S&P 500 by 8 percentage points over the past year. RC portfolio Performance | As of 11th Nov 2026 Now let’s discuss a few interesting ideas on our radar. Below, we present 5 potential rebounds. For each name, we cover why the stock is down, our take ( where we would get interested/consider buying, if at all ), and the KPIs we will track. Welcome to Rebound Capital. We study beaten-down stocks and businesses that have made successful comebacks. Subscribe for free to join our growing community of 28,800+ subscribers and make sure you don’t miss our next briefing. Subscribe now TEXAS PACIFIC LAND ($TPL) Texas Pacific Land is one of the largest landowners in Texas. They own ~870k acres of land in West Texas. Most of this land is in the Permian Basin, which is the USA’s busiest oil field. The business model is to lease its land to oil drilling companies and take a royalty on each barrel, as well as for other activities on its land, such as pipelines. Its total revenue was $798M in 2025, reported in 2 segments: Land ( $491M ) and Water ( $308M ). Its key business segments are: Oil and gas royalties ( $412M, about half of revenue ) Water sales ( $170M ) to fracking companies Water royalties ( $124M ): fee on the salty wastewater that comes up with the oil Easements and other ( $92M ). Rent from data centers being set up on its land would be a fifth revenue stream and is close to 0 today. The business is highly profitable and converts 60% of revenue to net profit. It needs almost no capital, so $498M ( ~62% of revenue ) became free cash flow in 2025. It has no debt. Moat The moat is the large land bank it owns in the Permian Basin. Nobody is building another large land bank like this in such a strategic area. Anyone who wants to drill, lay a pipe, or move water across it has to pay. The land bank is now valuable for data center construction, though that hasn’t materialized in 2-3 years of AI expansion. What Went Wrong TPL is down ~50%. The stock peaked at ~$547 in February. It now trades at ~$275. The drawdown is primarily due to TPL’s high starting valuations and high expectations for a material data center to be built on TPL’s land. Investors were too bullish on data center prospects: In December 2025, TPL invested $50M in Bolt, a data center developer chaired by former Google CEO Eric Schmidt. After this news, TPL’s stock rose ~80% in 2 months. This optimism has not translated into business results, though, as no material data center deals have been announced. Texas paused data center permits: In September, the outgoing governor told regulators to stop issuing data center permits until audits of the grid and water supply are done. A new governor takes office in January, and TPL’s management expects a similar approach. This is a live risk to the whole DC thesis in TPL. Doubts about the AI buildout: The market has become much more selective regarding the AI buildout. The market is demanding a certain ROI and waiting for projects to receive funding before assigning a multiple to expected AI revenue. Our Take This is an amazing royalty business model, with a valuation floor due to the large land bank ( though much lower than today’s price ). Our issue with the stock is its trading price. At ~$275, TPL is a ~$19B company trading at ~30x free cash flow. For a royalty business tied to oil, this is a very high multiple and reflects strong growth from data center projects on TPL’s land. A key risk is that the DC projects may not be as lucrative as the market expects. LandBridge, a peer, expects $10M-$20M a year for each GW of data centers on its land. Our base case assumes 3 to 5 GW gets built on TPL’s land by 2031. That would bring in ~$30M-$100M a year. Even the top of that range is small next to the ~$620M of free cash flow TPL makes today. Potential Buy levels and KPIs we are tracking We estimate the stock’s base value is ~$190 per share. We would get interested below ~$170. The most important KPI for this thesis is revenue outside oil and gas royalties ( water, easements, and data centers ). It grew 13% in Q3, with the revenue from data centers at 0. The current valuation already assumes a major data center deal. BOSTON SCIENTIFIC ($BSX) Boston Scientific is one of the world’s largest medical device companies. It is best known for its cardiovascular health products. The company has a deep moat. Doctors train for years on one company’s devices and rarely switch, and every new device needs years of clinical trials before it can be sold. Revenue was $20.1B in 2025 ( +20% YoY ), and ~64% of it came from the US. Cardiovascular is about two-thirds of revenue ( $13.3B ). Two products have driven the stock over the last 3 years. Watchman is an implant that lowers the risk of stroke in patients with an irregular heartbeat. Farapulse treats that irregular heartbeat with electrical pulses in place of heat or cold. Recently, growth in both of these growth drivers has collapsed, and in the most recent Q3 earnings, revenue was $5.14B, up just 1.5% YoY. The stock has lost its premium valuation primarily due to weaker growth. What Went Wrong The stock peaked at ~$108 in Sep’25 and now trades at ~$45, a ~59% drawdown. That is ~14x this year’s guided earnings. Three guidance cuts in one year: management has repeatedly ( 3 times ) cut its full-year revenue guidance. Originally, in February, management guided to 10.5%-11.5% sales growth for 2026. Recently, it has now guided for ~4.25% revenue growth for 2026. The EPS guidance has come down as well. Segments driving the last 2-3 years of growth are now shrinking: Watchman and electrophysiology products drove much of Boston’s growth. In the latest quarter, Watchman sales fell 6% YoY, and electrophysiology sales fell 5%. This is driven by increasing competition in these products and the consequent market share loss for Boston. It is very difficult to predict when this share loss can be stemmed. Weak 2027 commentary: Management has called 2027 the ‘trough year’. This implies another weak year for Boston. Given this commentary, the market will not re-rate the stock. Cyberattack in August: The breach cost $150M in revenue in Q3 and will also hit Q4 results. This is a temporary blip for us, but it did lead to poor sentiment regarding the company. Penumbra deal: Boston is taking on ~$11B of new debt to buy Penumbra for $14.5B. The deal closes in early 2027. Our Take The company has been losing share in its 2 blockbuster products, and it has no revenue-floor guidance for them. Until these products stabilize, the market will not re-rate the stock, and we will monitor the business results until then. The rest of the business is doing relatively well, and the company will be back on the growth path once the shrinking businesses stabilize. The next version of Farapulse is due in early 2028. Potential Buy levels and what we will track We will wait. We value Boston at ~$40 a share, about today’s price. The number we track is US electrophysiology sales growth, and it needs to stop falling before we would buy. The next date is the Q4 results and the 2027 guidance in early February. Next up, we present our research on Tyler Technologies, Deckers, and Marvell. Read more
View original →Deep Dive: $ASML
Grab a cup of coffee for this one. Arguably, this is the most in-depth research on ASML you can find anywhere :) The chip in the phone you’re reading this on is very likely made with machines from a single Dutch company. So is every advanced chip in the world. This company is ASML. ASML is the only company that can manufacture EUV lithography machines. Each of these sells for ~€230m, and without them, no AI chips can be manufactured. Even at this price, ASML’s customers can’t get machines fast enough, and the company can't meet demand through 2028. ASML's journey is one of the best business stories, and it nearly didn’t happen. ASML started working on EUV in 1997. 15 years later, despite the best efforts of some of the smartest engineers in the world, the technology was still years behind schedule. That year, the smartphone boom drove a sharp rise in chip demand, and ASML's largest customers decided they couldn't wait any longer. Intel, TSMC and Samsung bought 23% of ASML for €3.85bn and committed another €1.38bn to fund its research. Think about how unusual it is - multi-billion-dollar companies don’t usually fund their suppliers' research. They shop around for alternatives. But Intel, TSMC and Samsung all knew there was no alternative to ASML. Even with funding and technology support from some of the world's most advanced tech companies, the first EUV-made chips didn’t ship until 2019. It took nearly 20 years and ~$9bn in R&D before EUV chips went into production. But that persistence paid off. Today, the most advanced chips in the world, including the ones that run AI models, are printed on ASML’s EUV machines. No other company has brought one to market. To put the demand in context, ASML is booked out till 2028, and customers are asking for allocations in 2029 and 2030! Welcome to Rebound Capital. If you are new here, we conduct in-depth research on beaten-down stocks and study companies that have made successful comebacks. Subscribe for free and join 26,600 other investors so you don’t miss our next briefing. Subscribe now Check out our previous coverage of ASML. We updated the research in March 2026 and estimated ASML’s intrinsic value. Here is a quick order in which we will analyze the business: How an EUV machine works What ASML sells ASML’s moat and Chinese competition Demand drivers Why forecasts keep coming in low Valuation and intrinsic value Key risks What we will track Our take and rating Let’s dig in. How an EUV machine works An EUV machine has close to 100,000 parts. The system is so large that it ships in 3 cargo planes. This huge machine prints patterns too small to see. ASML EUV Machine The finest patterns or lines on circuits today are ~13nm wide. The latest high-NA EUV machines can even print patterns ~8nm wide. For comparison, a human hair is ~80,000 nm in width! One thing to note is that the industry’s node names don’t describe actual feature sizes. Nothing on TSMC’s 2nm node is 2nm wide. Transistor gates sit 48nm apart and the closest wires 26nm apart. The names are generation labels, and every chipmaker uses them that way. Lithography enables these small patterns on silicon. The key is that the smaller a circuit is, the more of them you can print on a wafer. This directly influences a chip's computing power and memory capacity. Nvidia’s Blackwell AI chip, made on an older TSMC process, already holds 208bn transistors across 2 dies. How it works: A lithography machine works like a projector. Light is sent through a stencil ( using this analogy to keep it simple ) of the chip’s design, called the mask. The light then hits a silicon wafer coated with a light-sensitive material. Patterns are created in the wafer due to the material reacting with the light. This process is then repeated multiple times layer by layer ( up to ~100 times ). Also, each layer must line up with the one below within a few nanometers. To draw smaller patterns, you need shorter-wavelength light. The older DUV technology used light with a wavelength of ~193 nanometers. EUV machines use light with a wavelength of just ~13.5 nanometers. This is what makes EUV so hard. There is no lamp or laser that gives off EUV light. So the machine has to generate its own. To do this, the machine uses molten tin. The machine melts Tin, then lasers (Trumpf is a specialized supplier for these lasers) hit 50,000 tin droplets a second, each about a third the width of a human hair. This superheats the tin droplets to temperatures many times higher than the Sun's surface. This creates a tin cloud/gas that emits EUV light. Of course, now you have to deal with the molten tin and clean it away. The next challenge is that almost everything absorbs EUV light, including air and glass. So the light travels through a vacuum, and the machine can't use lenses. Instead, it uses mirrors made by Zeiss. Each mirror has more than 100 layers, each only a few nanometers thick, and is smoother than an atom's width. If one mirror were the size of Germany, its biggest bump would be ~0.1 mm high. The mirrors are able to reflect a maximum of ~70% of the light. On top of that, light reflects off of upto 11 mirrors before reaching the wafer resulting in only ~2% of the light reaching the destination. This requires a powerful light source to compensate for the intensity loss. Then there is motion. The wafer must keep moving so the correct patterns are drawn. Since this printing happens at the atomic level, the wafer sits on a stage that starts and stops within a fraction of a second, and lands within a few nanometers of where it should be every time. Sensors check the wafer’s position up to 20,000 times a second. Any misplacement, even at the atomic level, is caught. This is only a glimpse of the complexity. EUV stacks one extremely difficult process on top of another, making it hard to compete. The capital intensity of the R&D process, and the effort required to create another Lithography competitor, makes it very likely that no other company in the Western world will compete with ASML in the future. Only China may eventually build its own EUV, but those machines will be locked out of the Western world. We see no threat to ASML’s monopoly. ASML’s key product segments ASML sells 2 kinds of lithography machines: Extreme ultraviolet ( EUV ) and Deep ultraviolet ( DUV ) EUV systems: Only ASML sells these systems. The chips they print run AI systems. The standard EUV machine ( called low-NA ) sells for ~€230m. The latest machine, called the High-NA, can print even finer patterns. It costs ~€350m and weighs ~165 tons! EUV system sales are estimated to account for a high-30s % of ASML’s 2026 sales. DUV systems: These older machines print most of the layers on every chip, including AI chips, and all of the chips in cars, appliances and most electronics. The most advanced DUV machines pass light through a thin layer of water to print sharper patterns, and ASML sells ~90% of them. Together with wafer-measuring and inspection tools, DUV sales are estimated to account for ~35% of 2026 sales. Installed Base Management: This is ASML’s service business. This section generates revenue from maintenance, upgrades, spare parts, and other customer services. One key element here is the upgrades business. Upgrades let customers get more out of a machine they already own ( for example, improving an EUV machine's efficiency ). In today’s environment, where new machines are booked out for years, customers pay for upgrades as they would for new capacity. This business is expected to grow 50% in 2026 and reach >25% of sales. In 2025, China accounted for ~33% of ASML’s sales, and that share is expected to fall to ~20% in 2026. ASML’s moat & Chinese Competition We think ASML probably would be on any top 5 list of companies with the widest moat in the equity world. The R&D moat: As we discussed before, the R&D for EUV took ~17 years and ~$ 9B. ASML’s own customers had to help with the technology. ASML spent ~€5B on R&D in 2025 ( ~14% of sales ), and Zeiss ( ASML has a ~25% stake ) alone holds more than 2,000 patents on High-NA optics. A rival would need to match decades of this spending before it sold a single machine. Getting skilled R&D researchers to work on this is not easy. Physics moat: Printing wafers at the nanometer level is not easy. EUV machines function at the edge of what’s possible. The light is made from using molten tin heated to temperatures up to ~40x the Sun’s temperature. The wafer stage starts and stops thousands of times an hour, landing within nanometers each time. ASML, Zeiss, and Trumpf have spent decades perfecting these individually difficult actions and then bringing them together in a working machine. This is not easy to achieve. Not in theory ( know-how ) or in the practical act of building such a machine. China’s prototype is EUV-capable but has not yet printed a working chip. The supply chain: ASML has an extremely complicated supply chain with >5,000 suppliers and close to 100,000 parts. Zeiss designs the machine's complex optics, and it has mastered this over decades, building its processes with ASML in mind. Similarly, Trumpf’s lasers and Cymer’s ( owned by ASML ) light source are built around ASML’s machines. Any rival would have to build this from scratch. Customer processes are designed around ASML machines: ASML’s customers, TSMC, Samsung, SK Hynix, Micron, and Intel, are racing to add capacity to manufacture silicon for the AI industry. Amid this rush, they have designed their current factories around ASML’s machines. Substituting the lithography machine in their manufacturing setup, when billions of dollars are on the line is an unacceptable risk. Even a competent competitor will need years and years of trials before anyone buys their machines. No chipmaker wants to focus on such non-core activities while AI demand is at a record high. Installed Base: Every machine ASML ships joins a fleet it services and upgrades for its lifetime. ASML gets real-world performance data, which lets it keep improving the machines. No rival will get any real-world data. How quickly can China build its own EUV machine? This is a key risk many investors cite. Many see this as the biggest threat to ASML’s monopoly. We will not brush aside these concerns, but from what we can see, China is some way away from building a comparable EUV machine. Even if China develops something comparable, the Western world won't use it. China will use it domestically to meet its own EUV demand, and as of today, no EUV machines are sold in China because of export bans. The real risk is how fast China can stop buying DUV machines from ASML. Or how many years will it take for ASML’s China revenue to go to zero? In our base case, this happens in the next 5-10 years. ASML's revenue gains from the AI revolution will overshadow this loss. Currently, China does have a state-owned company, Shanghai Aishengna, making DUV machines for the 28nm node. This is where ASML was about 15 years ago. Chinese commentators themselves say it lags ASML by about four generations . Reports say ~5 of these machines will ship this year and ~20 in 2027. ASML will ship ~130 immersion machines in 2026 alone. China’s machines still depend on imported parts, too. An unverified 2025 story claims a Chinese firm reportedly took apart an older ASML DUV machine to reverse engineer it. They tried to break it apart and reverse-engineer its features but couldn't put it back together, so they had to call ASML’s engineers. The story is unconfirmed ( it comes from a single US commentator ), but it is a fun reminder of how hard these machines are to copy. Demand drivers The biggest risk to any AI investment today is timing. We are in the third year of the AI supercycle, and the question is whether it is nearing its end. Capex digestion could stall spending, or even cause it to fall, over the next few years. The bear cases are well known, from low returns on investment to outright capex overspending. Here is what we see in the overall AI trade. The world’s largest and best-run firms say demand is voracious, and they are betting hundreds of billions of dollars on it. Use cases are proliferating ( Meta’s Muse has found product-market fit for consumer agents ), and open-source models are so good that companies can unlock significant productivity while spending far less than they did with frontier models. This bodes very well for AI infrastructure. The frontier labs may be in trouble. But that’s good for overall AI adoption. Having seen how well AI is doing in math, investment research, coding, and many other fields, we at RC are willing to bet we are in the middle of the AI revolution. Now we will present the demand signals for ASML’s business. All signs point to strong revenue and profit growth in the coming years: Capacity is sold out through 2027 and 2028: ASML will sell out all the units it estimates it can produce in 2027-2028 and cannot supply full demand. This also gives it tremendous pricing power. ASML is increasing capacity by 30% in 2027 and 30%+ in 2028: management said that even with a 30% increase in both years, supply will still not meet demand. They may increase the capacity by more than 30% in 2028. Such strong words from a generally conservative management team show how strong the demand cycle is. Customers interested in capacity for 2029 and 2030 today: customers want to lock in capacity early. Memory supercycle and incoming capacity additions: ASML expects its memory system sales to rise by 75% in 2026. AI memory chips need high-bandwidth memory ( HBM ), whose manufacturing requires more EUV lithography. The latest HBM chips actually use more EUV than previous generations. TSMC capex and revenue guide: Although TSMC did not guide to next year’s capex in their recent earnings call, its guidance is for AI accelerator revenue to grow at a mid to high 50s % CAGR from 2024 to 2029. The CEO said: ‘Whether it is a high 50 CAGR, it does not matter. It probably will be higher.’ We have built our bottom-up estimate for WFE capex using this update. ASML’s sales follow what its customers spend on new fabs. So we built our own estimate of what its 5 biggest customers will earn and spend out to 2030. Here is what we expect ASML’s 5 biggest customers to earn and spend ( $B ): TSMC: Its capex jumps from $41bn in 2025 to $60-64bn this year, and brokers expect $81-90bn in 2027. TSMC says its spending over 2026-28 will be significantly higher than in the past 3 years. We expect ~$115bn by 2030. That would be ~30% of its revenue, down from ~35% this year, because TSMC’s CFO expects revenue to grow faster than capex. Memory: Memory prices have risen sharply, and SK hynix’s revenue should more than triple this year. Samsung, SK hynix and Micron spent ~$70bn on capex in 2025, and we expect ~$164bn in 2027. We assume memory prices fall back in 2028, but capex keeps growing because much of today’s spending goes into buildings. Micron says its construction spending is growing much faster than its tool spending. The machines for the new fabs ( Yongin in Korea, and Micron’s fabs in Idaho and Japan ) go in during 2028-29. Intel: After cutting capex to $18bn in 2025, it plans to spend more than $20bn this year and “significantly” more in 2027 as it prepares its 14A process for 2028. We have ignored other potentially major spending sources, like SpaceX’s terafab and other large investments to secure sovereign AI manufacturing across the Western world. Is the valuation attractive after a ~2x upmove since Jun’25? ASML’s shares are up ~2x since our June 2025 report. Yet the stock trades at ~25x next year’s earnings today, below the ~28x it traded at back when we first wrote about it. Earnings estimates rose faster than the share price. This is what we like most about owning a business like ASML. The value it creates can outrun even a big stock move. Next, we dig into why the market is giving us this opportunity and why we think the market’s forecasts are too low. Then we present our Excel model and ASML’s intrinsic value estimates. Read more
View original →Hindsight Capital
Lululemon's weak product line is an internal problem warranting caution before buying.
View original →Industry Deep Dive: Semiconductor Wafer Fab Equipment
In this article, we share an industry primer on the machines that enable the creation of semiconductor wafers ( the silicon on which GPU/CPU chips are ‘printed’ ). The companies that sell these machines are few ( ~5 major ones ) in number and underpin the entire AI industry. This enables them to generate superior returns on capital. Roughly $150 billion will be spent on wafer fabrication equipment in calendar 2026, according to our estimates, and three-quarters of it will pass through these companies. We cover how a chip is made, where each dollar of equipment spending goes, and who collects it. We then value the four listed names we follow. We set out the levels at which we would be interested in these stocks. Though this industry is cyclical, spending on wafer fab equipment/machines has grown secularly over the last 30 years. With AI, this growth has exploded. Unlike memory, these companies are not at risk of commoditization with one another due to the greater complexity of the wafer fabrication process. The real competitive threat comes from China. Chinese toolmakers are taking share at mature nodes, which is one reason our base cases assume WFE companies will lose share there. But as the leading edge is growing so quickly, the net effect will be a gain in share. Welcome to Rebound Capital. If you are new here, we conduct in-depth research on beaten-down stocks and study companies that have made successful comebacks. Subscribe for free and join 26,600 other investors to make sure you don’t miss our next briefing. Subscribe now Despite strong share price gains in the last 1 year, these firms have recently corrected 15%-40%. In case of further drawdowns, we will be interested in these companies because: They are the cleanest way to own the artificial intelligence build without having to pick which chip wins. Our base case projects spend of ~$204 billion by 2030, and our bull case is even higher. They have a monopoly or duopoly position in an industry that is growing directly as a result of increased spending on AI. The key 5 companies in this sector are: ASML: makes lithography machines and has no competitor in the most advanced ones Applied Materials: Applied Materials sells into more process steps than any other company. It leads in the deposition process Lam Research: leads in the etch process and is the one most exposed to memory KLA Corporation: has roughly half the market for the tools that inspect and measure defects in wafers Tokyo Electron: Tokyo Electron is a Japanese firm with strong presence in deposition Every GPU and CPU begins as a silicon wafer. The equipment companies are the ones that turn that blank wafer into a working chip, which puts them at the base of the AI industry. A semiconductor wafer Wafer fabrication process Below, we can see the various steps involved in treating a pure silicon wafer to create the patterns and connections needed for electronics. This is a highly simplified diagram, and the process is not linear as shown below but circular. Each of these steps will be done multiple times ( even 100s ) to create the GPU or CPU. Source: Wafer Fab Process Leading edge chips are built in layers printed one on top of the other. This is a very difficult process as the dimensions are a few nanometers wide. Real life depiction The raw materials for making chips are discs of pure silicon, 300 millimeters across and polished flat. Deposition Every layer on the chip is started by coating it with the required materials. It may be a few atoms thick metal layer or some other electrical part needed in the chip. Lithography: drawing the pattern The lithography machine uses light to project patterns and define the circuits on the wafer. ASML’s latest machines can cost upwards of $350M per machine. Lithography sets the pace of the fabrication process as it prints the patterns on which other machines coat, carve, etch or finally check the circuits. Etch: taking material away This machine removes material from parts where the pattern has not been made. This is exceedingly difficult. For example, cutting a hole straight through 100+ layers without affecting the whole structure. Cleaning, treating, and polishing Cleaning is important as even a speck of dust can render the chip useless. Polishing is required to ensure the surface is even, as you can’t print the next layer correctly on an uneven surface. Metrology and inspection: checking the work Once the above processes have been performed multiple times and we have a wafer with multiple chips on it, it needs to be checked. This is very important as each wafer can be worth 100s of thousands of dollars at this stage. So, searching for defects is important because it improves the process yield. Packaging: increasingly important for AI chips Assembly used to be cheap work. AI changed that because today’s AI chips ( GPUs/CPUs ) are several chips stacked and wired together. This stacked 3D structure requires expensive packaging. In fact, packaging capacity is a key bottleneck. The spend here goes to smaller companies like BESI. We will not cover them in this article. Who buys wafer fab machines? The value chain works like this: Hyperscalers ( cloud providers and Meta ) and model labs order GPUs and CPUs from chipmakers like Nvidia, AVGO, Intel, and AMD. These companies own the intellectual property for chip design but don’t have their own manufacturing capabilities ( except Intel ). These companies then place manufacturing orders with TSMC, Samsung, and Intel ( mostly placing their own orders ). These companies manufacture the wafers and circuits physically. Inside their factories, they require Wafer Fab Equipment ( WFE ). Note: Capital spending estimates vary by estimation method The conversion ratio of Semiconductor capital spending to wafer fab equipment spend is roughly ~58%. The leading equipment companies The diagram below shows RC estimates of how $1 of capex for wafer fab equipment is allocated across process steps. Lithography is the largest share of spending at 28%. Within Litho, ASML is the large monopoly player. Then, the Etch and Deposition steps account for ~21%-22% of WFE spend. In these 2 steps, Lam Research ( LRCX ) and Applied Materials ( AMAT ) are the leading providers, with TEL also having a decent share. Process control is ~11%. The remaining ~18% of the spend covers ion implantation, polishing, cleaning, and coat-and-develop tools. The market share by company in each step is shown below: ASML is the monopoly player in Lithography. The 90% share shown above is for the entire lithography space. In cutting-edge equipment, ASML is the sole provider. Lam Research leads in Etch and is the most exposed to memory companies. AMAT leads in deposition and KLA in the process control step. In fact, KLA holds more than a 50% share in certain leading-edge process control steps. The key reason for sharing this report with you is that the WFE companies have brilliant business models. They have dominant market shares in critical equipment. As and when these stocks experience larger drawdowns, we will share our full deep dives on them. For now, we will share preliminary valuations for each company. RC WFE projections through 2030 Equipment spending was ~$18B in 1995 and ~$116B in 2025, using estimates from SEMI ( a semiconductor industry body ). That’s a nearly 6.5% CAGR over 30 years. The last decade’s CAGR was ~13% as the leading edge has become much more expensive. In 2026, the WFE companies are talking about strong growth exceeding 20%. Below, we show the RC estimate of WFE spending through 2030 for the bear, base, and bull cases. We have done the WFE forecasting bottom-up. We took the expected capex from Samsung, TSMC, Micron, SK Hynix, etc., and then forecast it through 2030. Then we applied a conversion ratio to that spend to estimate our WFE. The key difference among the bear, base, and bull scenarios lies in the shape of the capex. In our bull case, we expect AI spending to continue scaling, driving secular growth through 2030. In the bear scenario, we expect 2028 to be the year when WFE spending declines sharply, as manufacturers have overinvested in capacity. Our base case assumes secular growth, but 2029 WFE spend will be flat to marginally lower than 2028. Cyclical Nature of the WFE industry Historically, this has been a cyclical industry. We expect that to continue. We don’t believe those who say semiconductors have ceased to be cyclical, as they have no data to support it. The recovery from any decline in WFE spend may not be quick. This is the key bear pointer for us to watch if WFE companies enter a deep drawdown, which could happen if AI growth rates plateau or even just slow from their current high. Valuation of LRCX, ASML, AMAT, KLAC Read more
View original →September 2026 Portfolio Update
Nvidia's supply-constrained ~70% FY28 growth guide signals AI capex strength through end of 2027.
View original →