Melvin
@MelvinInvests • 35,146 subscribers
AI Analyst @MilkRoadAI | Finding opportunities across AI, photonics, defense, space, and tech.
Shorts
Videos
$NBIS is my #1 AI stock pick and I think Wall Street is completely underestimating what this company is building. I joined Charles V Payne on Fox Business to discuss why I'm so bullish on Nebius and where I see the company heading. What makes Nebius so interesting is that they're building an entire vertically integrated AI cloud platform, from the physical infrastructure and computing hardware all the way up to the software developers need to train, deploy, and run AI models. Think about what Amazon did with AWS. Instead of simply providing servers, Amazon built an entire cloud ecosystem around its infrastructure. Nebius is pursuing a similar opportunity, except its platform is being built specifically for AI from the ground up. And what I really like is how aggressively management is expanding its software capabilities. They acquired Tavily, which provides real time search infrastructure for AI agents. This allows developers to build AI systems that can access live information, conduct research, and interact with the internet instead of relying entirely on information stored inside a model. They also acquired Eigen AI, which specializes in making AI models run faster and more efficiently. As inference demand explodes, reducing the computing power and cost required to generate AI responses becomes increasingly important. Then there's Clarifai, where Nebius brought in its core engineering and research team and licensed its inference technology. While Eigen focuses on optimizing individual models, Clarifai's technology helps optimize how those models operate across the broader computing infrastructure, including orchestration, deployment, and resource efficiency. All of these technologies feed into Nebius Token Factory, its managed inference platform. Nebius wants to control and optimize the entire process, from the GPUs powering AI models to the software serving those models and the tools AI agents use to perform tasks. And the economics behind this will be enormous. The more efficiently Nebius can operate its infrastructure, the more computing capacity it can deliver from the same hardware. That potentially means lower costs for customers, better utilization of expensive GPUs, and stronger margins as the business scales. I think investors are still looking at Nebius primarily as another AI data center company, when the bigger opportunity is that it's positioning itself to become a full-stack AI cloud provider. The company is still in a capital-intensive expansion phase, and execution will matter enormously. But I believe the combination of infrastructure, proprietary software, strategic acquisitions, and rapidly growing AI computing demand makes Nebius one of the most compelling long-term opportunities in the AI sector.
Melvin91,767 views • 10 days ago
Wall Street is getting $CRWV COMPLETELY WRONG. I went on Making Money with Charles V Payne today to explain why I think this selloff is being blown WAY out of proportion. CoreWeave closed down 15.5% today at $65.78 and a lot of that panic is coming from its bonds. The company's 2032 bonds are now trading below 80 cents on the dollar, pushing yields close to 15%. For context, CoreWeave issued those same bonds just four months ago at 9.625%. Now, I think a LOT of people are misunderstanding what this actually means. First, CoreWeave is NOT suddenly paying 15% interest on those bonds. The interest rate is locked in at 9.625%. The 15% yield reflects what investors are willing to pay for that debt today and it gives us an idea of how expensive similar unsecured borrowing could be going forward. And there's a much bigger issue here that people are overlooking. Inflation came in hotter than expected this morning, and the 10 year Treasury yield is sitting around 5.2%. Think about it. If investors can earn over 5% lending money to the US government, why would they take on substantially more risk lending to a company like CoreWeave without demanding a MUCH higher return? That's what we're seeing across the bond market right now, and it's not just a CoreWeave problem. But here's where I think the market is getting this wrong. CoreWeave just raised $4.2 BILLION at 2.875% three weeks ago. Its Q2 revenue came in at $2.58 billion, more than doubling from a year ago, and its backlog is sitting at $104 BILLION, with another $25 billion in commitments added during early Q3. We're talking about a company with customers like $MSFT, OpenAI, and $META. The demand is there and the contracts are there. The biggest question is how cheaply CoreWeave can finance the infrastructure needed to fulfill those contracts. And this brings me to $NVDA. I genuinely believe people are underestimating just how important these neocloud companies are to Jensen Huang. Nvidia invested $2 billion into CoreWeave earlier this year and has a $6.3 billion agreement to purchase unused capacity through 2032. Why does that matter? Because Nvidia NEEDS these neoclouds to succeed. The hyperscalers are already developing their own custom AI chips to reduce their dependence on Nvidia. Neoclouds help Nvidia expand its customer base and create another massive distribution channel for its GPUs. Now imagine what happens if these companies suddenly struggle to finance their expansion. That's a massive problem for Nvidia, and I find it very hard to believe Jensen would simply sit back and watch one of his most important customers struggle without exploring ways to support it. Now, that doesn't mean there aren't risks. My biggest concern with the neocloud trade has always been dilution. If these companies can't borrow at attractive rates, they may have to issue more shares to fund their expansion. And that's something existing shareholders absolutely need to pay attention to. It's also why I think investors need to be selective. Not every neocloud has the same balance sheet, customer commitments, or access to financing. but I don't think the AI infrastructure thesis has suddenly changed because CoreWeave's bonds sold off. The demand for compute is still enormous, and these companies are playing an increasingly important role in meeting it. Short term, interest rates are driving the trade but long term, I still believe the AI infrastructure buildout is one of the biggest investment opportunities of our lifetime.
Melvin224,913 views • 27 days ago
We're only in the second or third innings of this massive AI infrastructure buildout and we're not going to see the full profitability of these investments until the eighth or ninth inning. One of the biggest concerns on Wall Street right now is the enormous amount of capital being poured into AI. Companies like Microsoft, Amazon, Google, and Meta are spending and raising hundreds of billions of dollars building data centers and I fully expect some of these companies to experience negative free cash flow for the next couple years as they continue ramping up their capital expenditures. But I don't necessarily think that's a bad thing, especially when you consider the scale of the opportunity ahead. We're essentially watching the construction of an entirely new computing infrastructure, and building that infrastructure requires enormous upfront investments before the full economic benefits become visible. Think about it. We're still in the earliest stages of AI agents, autonomous systems, enterprise AI adoption, and advanced reasoning models. As these technologies become more capable and widely adopted, the demand for computing power will increase dramatically. The important distinction is that the spending happens today, while much of the potential profitability may not materialize until years down the road. That's what makes this investment cycle so difficult for Wall Street to evaluate. Investors are looking at declining free cash flow and questioning whether these companies are spending too aggressively, while the companies themselves are preparing for demand they believe will be substantially larger in the future. Of course, there's no guarantee that every company will earn an attractive return on these investments. Some will overspend, some will struggle with execution, and others will emerge as major winners. But I think it's a mistake to judge a multiyear infrastructure buildout solely by the negative cash flow we're seeing during its earliest stages. Thanks for having me on the show, Charles V Payne! Always a pleasure discussing the markets and the incredible opportunities ahead in AI.
Melvin55,500 views • 9 days ago
Everyone needs to STOP worrying so much about this neocloud selloff. I think people are forgetting just how important companies like $CRWV and $NBIS are to the entire AI ecosystem, especially $NVDA. And remember what Jensen Huang announced just TWO months ago? Nvidia partnered with BlackRock, Blackstone, Goldman Sachs, Apollo, Brookfield, and KKR to help mobilize over $500 BILLION in financing for AI infrastructure. Think about what that actually means. Jensen understands that the biggest bottleneck in AI isn't necessarily demand anymore, it's getting enough capital, power, and infrastructure to actually meet that demand. Also, Jensen NEEDS these neoclouds to succeed because they give Nvidia leverage against the hyperscalers. Companies like $MSFT, $AMZN, and $GOOGL are developing their own custom AI chips to reduce their dependence on Nvidia over time. And that's exactly why neoclouds are so important. Companies like CoreWeave and Nebius are building enormous amounts of GPU infrastructure centered around Nvidia's technology. They give AI companies another place to access compute without having to rely entirely on the traditional hyperscalers. The stronger these neoclouds become, the more alternatives customers have and the less dependent Nvidia becomes on a handful of hyperscalers controlling AI infrastructure spending. Now imagine what happens if these neoclouds suddenly struggle to finance their expansion, which is exactly what we're starting to see with $CRWV. CoreWeave's bond yields just surged to over 15%, which is a warning sign. For context, these bonds were issued at a 9.625% yield just four months ago and now investors are demanding significantly higher returns to hold that debt. Essentially, the bond market is telling us that investors are becoming increasingly concerned about CoreWeave's financial risks, which could make future borrowing a lot more expensive. And that's a massive problem for $NVDA because these companies are buying billions of dollars worth of GPUs and helping expand the market for Nvidia's technology. This is why I believe Jensen will step in to support strategically important partners like CoreWeave if financing conditions become difficult. Nvidia has already demonstrated that it's willing to put capital behind this ecosystem. I'm not saying Jensen will bail out every neocloud, but I find it very hard to believe he'll just sit back and watch one of his most important customers struggle to finance its expansion, especially when Nvidia needs these companies to compete against the hyperscalers.
Melvin154,793 views • 27 days ago
I sold the majority of my $INTC position but I genuinely want Intel to succeed. I joined Charles V Payne on Fox Business today to discuss why I've lost confidence in Intel's turnaround following its latest earnings report. I had a tremendous amount of hope for this company, especially after Lip-Bu Tan took over as CEO. And to be fair, I still believe he's making some of the right decisions. But after listening to the earnings call, I realized that Intel's turnaround is going to take significantly longer than I originally anticipated. My biggest frustration is that Intel has one of the greatest opportunities in its history sitting right in front of it, and the company simply isn't positioned to take full advantage of it. We're witnessing one of the largest infrastructure investment cycles in history. Hyperscalers are spending hundreds of billions of dollars building AI data centers, demand for server CPUs is accelerating, and Intel's data center business is finally showing meaningful growth. Yet despite all this demand, Intel doesn't have enough manufacturing capacity to meet it. That's what makes this situation so disappointing. The customers are there, the demand is there, and the opportunity is there. Intel just can't produce enough chips to fully capitalize on it. Think about this because Intel generated nearly $79 billion in annual revenue back in 2021. Fast forward to 2026, during arguably the biggest semiconductor boom we've ever experienced, and the company is projected to generate just $63 billion. How does one of America's most important semiconductor companies end up generating LESS revenue during one of the greatest semiconductor investment cycles in history? And the problems extend beyond manufacturing capacity. Intel's PC business is facing pressure from higher memory prices, its foundry operations continue losing billions of dollars, and management isn't expecting meaningful improvements in supply until the second half of 2027. On top of that, Intel still isn't anticipating positive free cash flow next year. As much as I believe in the company's long term potential, I can't ignore those financial realities. There's also a massive difference between believing a company will eventually succeed and believing its stock is the best place to put your money today. Every dollar I have invested in Intel is a dollar I can't allocate toward companies that are already benefiting from the AI boom. That's ultimately why I decided to reduce my position so significantly but I genuinely want Intel to succeed, and I believe America needs Intel to succeed. Intel is one of the few companies with the technology and manufacturing ambitions to produce leading edge semiconductors on American soil. Today, much of the world's most advanced chip production remains concentrated in Taiwan, an island that China claims as its own. When you consider how dependent America's AI infrastructure, defense systems, cloud computing platforms, and broader technology industry are on advanced semiconductors, that concentration creates an enormous strategic vulnerability. We cannot build the future of American technology while remaining indefinitely dependent on overseas manufacturing for some of its most critical components. A competitive Intel would strengthen America's semiconductor supply chain, create more domestic manufacturing capacity, and provide an important alternative for companies looking to diversify where their chips are produced. I still believe Lip-Bu Tan has an opportunity to turn this company around. The engineering talent, manufacturing infrastructure, and strategic importance are all there. But Intel needs to execute, improve its manufacturing capabilities, control its spending, and demonstrate that it can translate the enormous demand in this industry into sustainable profitability. I'm not ruling out buying Intel again. If the company starts delivering meaningful improvements in manufacturing capacity, foundry economics, and free cash flow, I'll gladly revisit my investment thesis. For now, I'd rather watch the turnaround from the sidelines than continue betting such a large portion of my portfolio on promises that could take years to materialize. Appreciate Charles V Payne having me on Fox Business today to discuss all of this!
Melvin54,498 views • 11 days ago
South Korea's stock market is in FREE FALL and the panic surrounding the memory trade is getting completely out of control. The KOSPI just plunged another 7%, bringing its decline to roughly 40% from its June highs. Samsung and SK Hynix are getting absolutely destroyed, and the selling pressure is spreading across the entire AI infrastructure trade. I went on Making Money with Charles V Payne today to talk about the memory selloff and why I believe investors are losing sight of what's actually happening in this industry. And now we're getting ANOTHER warning out of South Korea. A new column published by Maeil Business Newspaper is raising concerns that the global memory industry could face oversupply beginning in 2029 as new manufacturing capacity starts coming online. The argument is that South Korea has become increasingly dependent on Samsung and SK Hynix, and that the country has roughly two to three years to develop other growth industries before the memory cycle potentially reverses. Now, I think this is something investors should absolutely pay attention to but here's where I disagree with the panic. We're talking about a potential oversupply problem in 2029 while the industry is still struggling to meet demand TODAY. Memory manufacturers are already dealing with capacity constraints, AI customers are securing long term supply agreements, and demand for HBM continues to grow as Nvidia $NVDA and other AI chipmakers push toward increasingly memory intensive systems. And people need to understand that building additional manufacturing capacity doesn't automatically mean we're going to have an oversupply problem. These fabs take years to construct, equipment needs to be installed, production yields need to improve, and manufacturers still have to qualify their products with customers. Not to mention, HBM manufacturing is substantially more complicated than traditional DRAM, and not all additional memory capacity can immediately be used to produce the most advanced AI memory products. Now, could we eventually see oversupply in 2029? ABSOLUTELY. Memory is a cyclical industry and I've never argued that these companies will enjoy record pricing and margins forever. But there's a massive difference between acknowledging that possibility and acting like the memory cycle is already over. I still believe the fundamentals of the memory trade remain incredibly strong, especially for companies positioned to benefit from HBM and the broader AI infrastructure buildout. Bullish on memory and if you want to see how I’m investing through this selloff and check out my full portfolio below.
Melvin81,464 views • 26 days ago
I’ve been saying for months that the majority of neocloud companies will NOT survive and Elon Musk just came out and said pretty much the same thing. There are simply way too many neocloud companies competing in an industry where the economics only make sense at scale. Building AI data centers requires billions of dollars in upfront capital, access to GPUs, massive amounts of electricity, and most importantly, financing. Now that debt markets are tightening, a lot of smaller operators are going to find themselves in a very difficult position. Some will get acquired, others will run out of money, and many will never become profitable businesses. However, I think the market is making a mistake by treating every neocloud the same. I’ve consistently said that I believe $NBIS, $CRWV, and $IREN will survive this consolidation, and here’s why. $NBIS is probably one of the best examples of what separates a serious AI cloud provider from the hundreds of smaller operators. Unlike companies that are building data centers and hoping customers eventually show up, Nebius already has more than $40 billion in long term customer commitments, relationships with investment grade customers like Microsoft and Meta, and access to substantial amounts of contracted power. Just a few days ago, Nebius secured another $3.5 billion financing facility from Citi, even as the broader market was questioning whether neoclouds could continue raising money. That alone should tell you something. Lenders are becoming more selective, but Nebius is still securing capital because it has actual customers and contracted revenue supporting its expansion. CoreWeave $CRWV is another company I believe will survive, despite my concerns about its debt and aggressive spending. The company reported approximately $104 billion in revenue backlog at the end of Q2, with another $25 billion in net customer commitments added early in Q3. It also has established relationships with some of the largest AI companies in the world and already operates infrastructure at a scale most competitors cannot come close to matching. Yes, its debt is a serious concern, especially as borrowing costs rise, but there’s a massive difference between a company financing infrastructure against committed customer demand and a smaller operator borrowing billions without knowing whether it can fill its data centers. Then there’s $IREN, which I think people are overlooking in this discussion. One of IREN’s biggest advantages is its power and infrastructure footprint. Securing electricity has become one of the biggest bottlenecks in AI infrastructure. IREN already controls a substantial pipeline of power- connected and developing data center sites, which would take competitors years to replicate. The company also has major contracts with Microsoft and NVIDIA, customer prepayments, and billions in GPU financing. That doesn’t eliminate execution risk, but it puts IREN in a much stronger position than operators starting from scratch. If the majority of neoclouds disappear, the demand for AI compute doesn’t suddenly disappear with them. That demand can shift toward the operators that have the financing, power, infrastructure, and customer relationships to actually deliver it. In other words, consolidation could eventually make the strongest companies even stronger. If the market is smart, $NBIS, $CRWV, and $IREN should actually move higher tomorrow following Elon’s comments. Obviously, given the current market environment, I wouldn’t be surprised if everything continues selling off together. But I think investors need to start separating the companies that are genuinely at risk of going out of business from those that are positioned to survive and potentially benefit from this consolidation.
Melvin33,006 views • 16 days ago
The majority of neocloud companies will NOT survive. I strongly believe we're going to see a MASSIVE amount of consolidation across this industry over the next several years, and only a handful of companies will come out on top. Personally, I think $CRWV and $NBIS will be the top 2. Just look at how competitive this industry is becoming. You have the traditional hyperscalers like Microsoft, Amazon, and Google spending hundreds of billions of dollars building their own AI infrastructure. Then you have companies like CoreWeave, Nebius, Crusoe, Lambda, and several others trying to capture a piece of that same market. As we know, building a neocloud is EXTREMELY capital intensive. We're talking billions of dollars in GPUs, data centers, power infrastructure, networking equipment, and everything else required to operate these facilities. And that's before you even consider the cost of financing all of this. Not every company is going to have access to the capital necessary to compete, and more importantly, not every company is going to have customers willing to sign multibillion dollar contracts. This is why I believe we're eventually going to see smaller neoclouds either go bankrupt or get acquired by larger players. I wouldn't be surprised if the industry eventually becomes dominated by just a few major independent GPU cloud providers. But here's why I'm particularly bullish on CoreWeave and Nebius. Both companies have already demonstrated that some of the largest technology companies in the world are willing to commit billions of dollars to their infrastructure. CoreWeave has built an enormous contracted backlog, while Nebius has secured major agreements with Microsoft and Meta. And then there's $NVDA, which I think is one of the most underrated parts of this entire conversation. Jensen Huang has a HUGE incentive to make sure the neocloud industry succeeds. Think about it. Why would Nvidia want the entire AI infrastructure market controlled by Microsoft, Amazon, Google, and Meta when it can help create another group of major customers buying billions of dollars worth of GPUs? We've already seen Nvidia make strategic investments across this ecosystem, including its investment in Nebius. Now, that doesn't mean Jensen is going to bail out every struggling neocloud, but I do believe Nvidia will continue supporting strategically important partners where it makes financial sense. And this brings me back to what's happening with CoreWeave right now. Investors are absolutely freaking out over its borrowing costs, and I understand why. Financing an enormous infrastructure buildout when interest rates are elevated is not easy, especially when you're already carrying significant debt. But I think people are confusing financing risk with demand risk. The demand for AI compute hasn't suddenly disappeared. If anything, the latest growth we're seeing from companies like Anthropic suggests we're going to need substantially MORE computing capacity over the next several years. That doesn't guarantee CoreWeave or Nebius will succeed. Both still need to execute, manage their capital requirements, and eventually demonstrate that their businesses can generate sustainable free cash flow. But I think the companies that survive this period could emerge in a MUCH stronger competitive position as weaker players get pushed out of the market. You can check out the full podcast here:
Melvin43,831 views • 26 days ago
Jensen Huang is investing in every photonics company he can find and the reason why tells you everything about where AI is headed (Save this). Lip-Bu Tan, the CEO of Intel says, when he looks for investment opportunities, he looks for the bottleneck and right now, the bottleneck is the interconnect, the pipes that move data between chips inside an AI data center. That is why he backed Credo Semiconductor, Astera Labs and Celestial AI on the optical side. Here is the simple version of what the interconnect bottleneck actually means. Think of an AI data center like a city, the GPUs are the buildings where all the work happens but for those buildings to function, you need roads connecting them, fast roads that can carry enormous traffic without congestion. And those roads are now the single biggest constraint on AI performance. As clusters scale to hundreds of thousands of GPUs, traditional copper wiring is hitting its physical limits and that is where this entire sector comes in. Credo Semiconductor (CRDO) is the most direct pure play on this theme, Credo makes high speed cables and optical chips that connect GPUs inside data center racks. Their revenue tripled in fiscal 2026 to $1.3 billion, growing 272% year over year at its peak and four of the world's largest hyperscalers each individually account for more than 10% of Credo's revenue. Astera Labs (ALAB) solves the connection problem between different chip types. Astera makes the PCIe and connectivity chips that manage data flow between GPUs, CPUs, and memory without errors or slowdowns. Their revenue grew 93% year over year to $308 million in Q1 2026 alone. The optical companies are where the longer-term and potentially larger opportunity lives. Copper has physical limits, you can only push electrical signals so far before the signal degrades, the heat spikes and power consumption explodes. The solution is light, fiber optic connections that move data using photons instead of electrons which is faster, cooler and far more energy efficient. Jensen Huang made this clear at Computex 2026 because copper works as long as physically possible but at greater distances and larger scale, optics takes over. Coherent (COHR) is the most established optical company in this space. Coherent makes the lasers, transceivers, and optical components at the foundation of all fiber optic communications. Nvidia signed a multibillion-dollar purchase commitment and invested $2 billion directly into the company and their customer order books are already extending out to 2028. Marvell (MRVL) is the most comprehensive bet across the entire connectivity stack. Marvell makes chips for optical networking, PCIe switching and custom AI silicon. Jensen Huang called Marvell the next trillion dollar company at Computex 2026 and backed it with a $2 billion Nvidia investment. Marvell also acquired Celestial AI, the exact company Lip-Bu Tan backed for $3.25 billion, gaining photonic fabric technology delivering 16 terabits per second of bandwidth. Lumentum (LITE), Corning (GLW), and Ciena (CIEN) round out the major public names. Lumentum received a $2 billion Nvidia investment for laser and photonics components. Corning known mostly for phone glass received $500 million from Nvidia for optical connectivity work and is up over 100% year to date. Ciena runs the optical networking systems between data centers and is seeing analyst price targets raised on the back of the AI optics boom. Every time a hyperscaler spends a billion dollars on Nvidia GPUs, the surrounding infrastructure, cables, switches, transceivers, optical components has to be upgraded to match. The smarter the GPU gets, the more the interconnect matters. Nvidia has committed at least $6.5 billion to photonics companies in the past 4 months alone and the companies building the roads between the GPUs may end up being just as valuable as the companies building the GPUs themselves. Follow me Melvin for more AI, semis and the next big market themes.
Melvin153,158 views • 4 months ago
The selloff in Micron is one of the best buying opportunities you'll see this year (Save this). Sanjay Mehrotra just explained exactly why the old mental model for Micron, cyclical, commodity, mean reverting no longer applies. Every AI system, regardless of what device it runs on, requires more memory at higher performance to unlock its full potential. From data centers to smartphones to autonomous vehicles, memory is no longer a supporting actor but rather the critical bottleneck determining how fast AI can move. What makes this cycle structurally different starts with what happened in 2023. Certain customers drove industry pricing to one third of 2022 levels, forcing Micron into severe losses while still requiring $10 billion in investment just to stay competitive. Most companies in that situation cut spending and survive but Micron invested through the pain with the vision that the other side would be worth it. Those 2023 investments are now producing 84.9% gross margins, $41.46 billion in quarterly revenue, and Q4 guidance of $50 billion up from $11.3 billion in the same quarter just one year ago. That is what it looks like when a company bets on itself at exactly the right moment. Even Micron's own largest customers, Nvidia, Google, Amazon could not forecast the scale of AI memory demand that materialized. When the biggest technology companies in the world cannot project their own memory requirements, you are watching a structural transformation that nobody had models to predict, still in its early innings. Supply cannot respond quickly enough to close that gap. Mehrotra confirmed on air that tightness extends beyond 2027, new domestic fabs take years to bring online, and new HBM capacity which requires advanced 3D stacking that compounds in complexity at every generation won't meaningfully arrive until late 2028. There is no fast fix to a shortage of the most valuable memory on earth. The strategic customer agreements are the most underappreciated part of the entire story. Multi-year contracts with volume commitments and price floors now cover roughly 20% of DRAM volume and 30% of NAND volume, locking in a $100 billion contractual revenue base. The old Micron was at the mercy of customers who could crater prices overnight while the new Micron has contractual floors that make the 2023 scenario structurally impossible to repeat. Long Micron and make sure to follow me Melvin for more deep dives into AI and memory.
Melvin131,779 views • 4 months ago
Dylan Patel just mapped out the most important investment theme in AI infrastructure (Save this). "In about two years, solar plus battery will be cheaper than gas." Every new NVIDIA Blackwell rack pulls 120 kilowatts, Rubin Ultra rack pulls 600 kilowatts and the next generation hits a megawatt. The US grid cannot keep up, interconnection queues now run five years in many markets so the entire industry is being forced to solve power from first principles. The solar thesis is already happening. BloombergNEF's 2026 LCOE report, covering 800+ financed projects across 50+ markets puts solar plus 4 hour battery storage at $57 per megawatt-hour. Combined cycle gas turbines hit $102 per megawatt hour, the highest on record, up 16% year over year. In California and parts of Texas, solar plus storage is already cheaper than gas for data center power today and solar panel costs are expected to drop another 30% by 2035. Getting power from the grid into the form chips actually require is an entire industry unto itself and NVIDIA just rewrote the rules. The 800 volt DC transition is the most important infrastructure shift that's happening right now. Today's data centers run on 48 volt DC power delivery, a single next-generation GPU pulls over 2,500 watts and at 48 volts, the current required to power a megawatt rack would melt the copper wiring. The investment thesis breaks into four layers and the first layer is power semiconductors, specifically silicon carbide and gallium nitride. At 800 volts, traditional silicon based IGBTs hit their physical limits. SiC and GaN devices are the mandatory replacement. Infineon estimates $175,000 of semiconductor content per megawatt of AI rack power, versus almost nothing today and by 2030, power semiconductor content per AI cabinet grows from $15,000 to $115,000+. The names here are Infineon ($IFNNY), ON Semiconductor ($ON), Wolfspeed ($WOLF), Navitas ($NVTS), and STMicroelectronics ($STM). The second layer is power management and conversion. Vertiv ($VRT) is NVIDIA's lead architectural collaborator for the 800V transition, building the hardware that converts grid AC to 800V DC and the DC to DC power shelves for ultra dense racks. Eaton ($ETN) and Monolithic Power Systems ($MPWR) round out this layer. The third layer is grid to site infrastructure, GE Vernova ($GEV) builds the heavy electrical equipment that connects utility power to the data center campus. Orders are running at twice the rate of shipments, the classic leading indicator of sustained multi year revenue growth. The fourth layer is behind the meter power generation like your bloom energy because grid interconnection queues run five years, hyperscalers are bypassing the grid entirely, building dedicated gas, solar and battery systems on site. Make sure to follow me Melvin for more opportunities across the AI supply chain.
Melvin108,541 views • 4 months ago
Micron is going to $4,000 and here is why (Save this). For 25 years, DRAM prices did one thing, they went down. Memory makers overbuilt, supply overwhelmed demand, buyers had all the negotiating leverage and that commodity trap crushed memory stocks every single cycle. What you are watching right now is a complete structural break from that 25 year trend. DRAM contract prices are up 700% year over year and the reason is AI and it is not going away. HBM3 was 12 layers, HBM4 in production and shipping now to Nvidia's latest GPUs is 16 layers. Each generation consumes significantly more wafer to produce than the last, meaning supply structurally tightens as the technology advances. Memory was 8% of hyperscaler capex in 2023 but is 35% in 2026 and is projected to hit 48% in 2027. Nearly half of everything Microsoft, Amazon, Google, and Meta spend on infrastructure will go to memory by next year. Going from the GB300 to the Vera Rubin 200 generation, GPU cost went up 57% while memory cost went up 435%. There are three companies on earth that can make DRAM at scale, Samsung, SK Hynix, and Micron. Both Samsung and SK Hynix are converting capacity to HBM which means conventional DRAM supply tightens further for everything else, and Micron captures pricing on both sides. Micron guided to $33.5 billion for Q3 and they reported $41.46 billion, a $7.96 billion beat, the largest earnings beat in the company's history. Gross margins came in at 85% above the 81% they guided. For Q4, they are now guiding to $50 billion in revenue with ~86% gross margins and $31 EPS. At $112 EPS in FY2027, the pre-earnings consensus and a 35x multiple, that is a $3,920 stock but with Q4 guiding to $31 EPS alone in a single quarter, FY2027 estimates will be revised meaningfully higher. Deutsche Bank says the supply-demand gap worsens through all of 2027 and into 2028. The market still thinks this is a cyclical bounce but this is far from it. This is the first chapters of a multi year repricing of the most critical component in the AI economy and Micron is at the center of it. Follow me Melvin for more AI, semis, and the next big market themes.
Melvin94,150 views • 4 months ago
Google just admitted it can't build data centers fast enough, so it's planning on baking its AI model directly into the chip instead (Save this). Google's new Frozen v2 chip permanently embeds parts of Gemini's architecture into the silicon itself, cutting down on the calculations and data movement needed to answer a query. Engineers estimate it could process 6 to 10 times more tokens per unit of power than Google's current TPUs. The real story is why Google is building this in the first place because Frozen v2 is meant to ease a severe internal compute crunch that's caused friction between teams at Google. It reportedly pushed Google Cloud to turn away outside business because it simply doesn't have enough spare capacity to go around. If Google, one of the largest chipmakers in the world, is short enough on compute to turn away paying Cloud customers, that's confirmation this shortage isn't a scaling problem unique to smaller players, it's systemic across the entire industry. This ties into a much bigger power struggle happening right now. More AI companies are trying to cut their reliance on Nvidia by building their own chips. OpenAI rolled out a custom chip called Jalapeno alongside Broadcom last month and Anthropic is now partnering with Samsung on something similar. The reasoning is pretty simple, Nvidia effectively acts as landlord for every hyperscaler out there, and the rent isn't cheap. Nvidia hardware can account for anywhere from 20% to 60% of total AI infrastructure spend, and once a company is tied into its ecosystem, every hardware refresh forces another costly one. That's exactly why Google built TPUs and Amazon built Trainium, both trying to protect their own margins for shareholders. Bullish on Marvell + Broadcom who makes these custom chips and follow me Melvin for more infrastructure plays and check out the link below for more!
Melvin63,767 views • 3 months ago
Nebius will be a TRILLION dollar company and here is exactly why (Save this). Brad Gerstner's Altimeter said on camera that they are invested in ClickHouse, and explained exactly why in one sentence: "If you're in the data infrastructure layer, then token consumption is driving a lot more consumption of your basic services." The flip side of that point is equally important. Gerstner added that the closer you are to a point solution, a single use app built on top of AI, "that feels like you're on the front of the conveyor belt heading toward the guillotine." Models get better, apps get commoditized and the companies that own the foundational infrastructure that every AI application must run through keep compounding. ClickHouse is exactly that foundational layer. It is a real time analytical database engine originally built inside Yandex, optimized for the exact query patterns that AI agents, LLM observability pipelines, and machine learning infrastructure generate, massive write volumes, complex aggregations, and sub-second response at scale. It processes hundreds of billions of rows per second, serves over 2,000 enterprise customers including Cloudflare, Uber and ByteDance, and grew 300% in a single year. In January 2026, a $400 million Series D valued ClickHouse at $15 billion more than double its $6 billion valuation just eight months prior. Here is where Nebius comes in. Nebius holds a 28% stake in ClickHouse, an asset that traces back to its Yandex origins. At ClickHouse's current $15 billion valuation, that stake is worth approximately $4.2 billion, sitting largely unrecognized on Nebius's balance sheet while most market coverage focuses entirely on the AI cloud business. A ClickHouse IPO, which the company is actively positioning toward, would force the market to mark that position to full public market value for the first time and could alone reprice Nebius meaningfully. But that hidden asset is just one layer of the bull case. The core AI cloud business just printed 684% year over year revenue growth, $399 million in Q1 2026 against $50 million a year prior. AI specific revenue grew 841% and now represents 98% of total revenue. The moat underneath those numbers is 3.5 gigawatts of secured power capacity, a $27 billion five year contract with Meta, a $2 billion strategic investment from Nvidia, and a Microsoft partnership ramping to full run rate in 2027, all stacked on top of a ClickHouse stake that the market is still not fully pricing in. Long Nebius and make sure to follow me Melvin for more underlooked AI oppurtunities.
Melvin69,634 views • 4 months ago
Nvidia’s CTO just explained why $NBIS stands out from every other neocloud. A normal cloud and an AI factory are not the same thing. A traditional cloud is built for flexible, general purpose computing. An AI factory has to make the GPUs, CPUs, storage, networking and software work together like one giant system. That is basically what Nebius is building. Nebius VP of Product Narek T. says customers are choosing the platform mainly because of reliability and performance. Having the newest Nvidia GPUs is nice but at this scale that is not enough. Customers care about how efficiently thousands of GPUs work together and, more importantly, what happens when something breaks and something will break. When you are running thousands of GPUs, hardware failures are inevitable. Nebius says its platform can automatically detect and replace unhealthy nodes in roughly 12 to 15 minutes while protecting the workload. Manually, that process can take hours. That matters because an AI lab does not want a massive training run getting interrupted every time one server fails. Nebius is also getting extremely early access to Nvidia’s newest systems. It was the first cloud provider in Europe to put GB300 NVL72 into production and the first globally to connect it with 800 Gbps Quantum X800 InfiniBand. Each GB300 NVL72 rack combines 72 Blackwell Ultra GPUs and 36 Grace CPUs into one liquid cooled system, with NVLink essentially making the rack behave like one enormous GPU. Nebius can then connect multiple racks together for even larger training and reasoning workloads. And they are already preparing for Vera Rubin.mThis is where the story gets more interesting to me. Nebius does not want to just rent GPUs by the hour forever. The company is also building serverless inference products where customers pay for actually running models and producing tokens and that could be a MUCH better business. Nvidia recently said Nebius software optimizations increased token throughput on the same GB300 hardware by as much as 2.7x in six months, while reducing cost per token by more than 60%.. That means Nebius can generate more tokens, more revenue and better economics from the same amount of data center power. Extremely bullish on $NBIS
Melvin21,381 views • 1 month ago
Goldman Sachs just quintupled its robot forecast and Morgan Stanley thinks even that is far too conservative (Save this). Goldman's old base case called for 1.4 million humanoid units in 2035 and the new one calls for 6.48 million. The 2030 estimate jumped from 256,000 to 890,000 units, putting the 2035 market around $138 billion versus $38 billion before. Morgan Stanley models a different universe entirely, with 24.4 million units installed by 2036 and 1 billion by 2050 worth about $7.5 trillion in annual revenue. The revision happened because the economics crossed a line because average unit prices are expected to fall from $41,800 in 2025 to $21,300 in 2035. Goldman sees e commerce warehouses as the earliest deployment, followed by auto production lines, with Amazon and Walmart leading. Amazon's automation program alone should save about $72 billion cumulatively from 2026 to 2030, adding roughly 240 basis points to group EBIT margin. The labor math explains why the long term numbers get so large because about 75% of US occupations and 40% of employees have some degree of humanoidability, implying roughly 63 million units and a $3 trillion market in the US alone. China is projected to hold about 302 million units by 2050 versus 77.7 million in the US, making this a geopolitical race too. Now here is some of the best positioned names sit one layer below the robot brands. Goldman named Toyota, Honda, Mitsubishi Motors, JTEKT, Aisin, MinebeaMitsumi, Renesas, Harmonic Drive, NEC, and Fujitsu as beneficiaries. Component and reducer makers like Harmonic Drive, Leader Harmonious Drive, and Shuanghuan carry buy or outperform ratings from Deutsche Bank, Bernstein, and UBS. Semis benefit directly, since each robot carries $3,000 to more than $6,000 of chip content across Nvidia, Renesas, and Teradyne. If you want ETF exposure runs through KOID as a broad physical AI basket and HUMN, which leans toward UBTech, Tesla, Harmonic Drive, and Hyundai. Suppliers are the cleaner bet here because no single robot brand will dominate and Goldman thinks Toyota alone could build 190,000 to 540,000 units in 2035, or just 3% to 8% global share. That fragmentation means demand for reducers, bearings, and actuators rises no matter who wins and Morgan Stanley notes the bar is now shifting to reliability, yield, and manufacturing scale, which favors proven incumbents. Bullish on humanoid robotics and especially the suppliers behind the buildout. If you enjoyed reading this, make sure to follow Melvin for more robotics and AI insights And if you want to see exactly what I'm buying as an analyst at Milk Road Pro, you can join for just $1 using the link below.
Melvin32,331 views • 2 months ago
People keep saying $MU can just build more fabs to solve the shortage but it doesn't work like that. AI demand is growing faster than the memory industry can add capacity, and new fabs cannot be built overnight. Micron’s first new Idaho fab will not begin producing wafers until mid-2027, its second Idaho fab and Japanese expansion are expected to start output in late 2028, and production at the New York site is not expected until 2030. Even after a fab opens, installing equipment, qualifying products and reaching full production take additional time. This means Micron cannot immediately satisfy every customer but neither can Samsung or SK Hynix. The entire industry is dealing with the same construction timelines and technical challenges, while HBM and advanced DRAM require more manufacturing capacity than traditional memory. Micron expects both DRAM and NAND to remain supply constrained through 2027 and 2028, with no clear timeline for when supply will catch up with demand. More than 75% of Micron’s 2027 output is already committed, and discussions are shifting toward 2028 capacity. Customers are now locking in supply years ahead of time because Micron has signed 26 strategic customer agreements covering an estimated 35% of its revenue through 2030. Those agreements include long term volume commitments, pricing frameworks and $32 billion of customer financial commitments, mostly through cash deposits. That completely changes the memory cycle because in the past, producers built too much capacity, supply exceeded demand and memory prices collapsed. Today, customers are reserving capacity in advance because they are more worried about securing enough memory than negotiating the lowest possible price. Micron does not need unlimited supply right now but rather needs controlled supply growth, strong pricing and customers willing to fund future capacity and that is exactly what is happening. Bullish on Micron and I’m still heavily positioned around the memory cycle, and if you want to see exactly how I’m positioned in $MU and the rest of the memory trade, check out my full portfolio below.
Melvin20,007 views • 1 month ago
Nvidia's next generation chips are about to make a handful of companies impossible to ignore and here is how you can benefit from this (Save this). Goldman projects networking content per AI compute system could rise from $315,000 today to $9.4 million in Nvidia's next generation, a 29x increase in what companies spend just wiring their chips together. Morgan Stanley, Citi and Goldman all agree on the same direction, projecting the total addressable market for AI networking to grow from around $11 to $15 billion today to $154 billion by 2028, roughly a 9x jump in just a few years. As AI clusters scale into the hundreds of thousands of GPUs, the chips themselves stop being the bottleneck and the connections between them become the limiting factor instead. Every GPU needs to talk to every other GPU almost instantly to keep a training run synchronized and that requires far more sophisticated switches, optical modules and cabling than today's networks can provide, which is exactly why Nvidia's next generation systems are expected to need dramatically more networking hardware per system just to keep up. Optical modules specifically are becoming the chokepoint. The global optical module total addressable market is projected to grow from roughly $6.7 billion in the first quarter of 2025 to nearly $69 billion by 2028, with 1.6T modules going from essentially zero market share today to capturing the majority of shipments by 2028 as speeds keep climbing. Nvidia sits at the center of this story since it designs the switches and networking silicon that go into every one of these systems, meaning that 29x jump in networking content per system flows straight into Nvidia's own revenue per AI factory, not just its GPU sales. Marvell is one of the biggest direct beneficiaries here, since it makes the custom networking chips and interconnect silicon that hyperscalers rely on to move data between GPUs at these speeds. Credo Technology plays a slightly different role, supplying the high speed connectivity chips that clean up signals moving through copper links, which becomes more valuable as networks push toward higher bandwidth with lower power draw. AAOI, or Applied Optoelectronics, sits closer to the physical optics layer, making the lasers and optical components that go into the transceivers carrying data across these networks, positioning it directly in the path of that jump from $6.7 billion to $69 billion in optical module demand. Coherent and Lumentum round out the optical side, both supplying lasers and photonic components that scale directly with 1.6T module adoption as it goes from near zero to the majority of shipments by 2028. Switch and connector makers benefit almost mechanically from this trend too, since every additional dollar of networking content per system has to physically pass through a switch, a cable, or a connector, meaning revenue scales with network complexity regardless of which chipmaker's GPUs sit inside the rack. Bullish on AI networking, make sure to follow Melvin for more semiconductor insights and check out the link below for more details.
Melvin42,220 views • 3 months ago