Are NeoClouds the Biggest Buying Opportunity Since Micron?

Think about what Jensen has done here in the last like 6 months. She has put 2 billion into CoreWeave, 2 billion into Nebius, equity in Lambda, N Scale, Crusoe. >> What's up everybody? It's LG DuBois here and welcome to Milk Road AI, the daily AI show that learns more about this stuff every single day and it hopes I can remember it next time I try to sound smart at a party. Today is July 14th, 2026 recording on the 13th. Neo Clouds, they are on a big dip right now along with much of the market, but many people are still very bullish on these companies getting a huge allocation of all that CAPEX that's coming in the next few years, including Melvin, our leading AI analyst here at Milk Road. Every week we sit down with Melvin to talk through one major part of the industry and the best looking stocks in that area and even get a little education on how it all works. And today we are going to talk about Neo Clouds, Nebius, CoreWeave and a lot of the other companies in the space, so tune in. And before we start, our analysts like Melvin made over 30 moves last month across their portfolios in AI, robotics and crypto. And this is after calling Micron, AMD, Bloom and many other AI winners earlier this year. You can see exactly what he's calling next for just a dollar in Milk Road Pro at the link below. Today's episode is brought to you by Bitget, Stocks 2.0 with real liquidity, real dividends and Securitize, the regulated rails for tokenization. Melvin, it's Monday. I'm ready for the alpha man. Your last couple videos have been total bangers. I'm learning a lot every time we do this and today we're going to talk about Neo Clouds, which is something where I like I it's one of those things I can be honest with you, man. I pretend like I know what that is and even after you told us about the dip a couple weeks ago, like I told a buddy IRL, I was like, "Hey, you got to buy the Neo Cloud dip." But on a honestly, I was like I I'm glad he didn't ask me any more questions cuz I was going to be like, "I don't know. You got to listen to Melvin to to explain this." And that's what we're going to do today. So I'll I'll let you take it away. >> Thanks for having me back on LG. Super excited to be here. I think last couple videos we talked about Micron, robotics. So today is all going to all going to be about Neo Clouds and specifically two to three stocks that I'm currently buying and I think these two are the most most interesting setups in the entire AI infrastructure right now. But before I can get into those stocks, I want to set the scene with some numbers that just dropped this morning from Morgan Stanley because they're directly relevant to everything we're going to be be talking about LG. So Morgan Stanley just raised its 2027 and 2028 CapEx estimates by 9 and 10%. And the new numbers are 1.23 trillion in 2027 and 1.4 trillion in 2028. Those figures now include SpaceX terrestrial compute CapEx for the first time as Morgan Stanley is now treating SpaceX as a legitimate hyperscaler class spender alongside Google, Amazon, Microsoft, Meta. So to put put all this in context, trajectory in context, hyperscalers have essentially spent 261 billion in 2024 and is now projected to spend 1.4 trillion in 2028, which is a 5x increase in just four four Yeah, in four years. >> Yeah. >> Now these numbers doesn't even include the CapEx for major frontier labs like OpenAI and Anthropic. And this is just Google Amazon Microsoft Meta and SpaceX. There's so many more companies that are spending billions of dollars like Micron. You have SK Hynix. You know, none of those are included in here. These are just the top five, you know hyperscalers. Yep. Now what does this mean? Like how can you benefit from all this, right? And one of the ways in which I think you can benefit is from through Neoclouds. Now if you can go to the next slide, is is a chart of Neoc Neoclouds. >> Wait, can I stop you Melvin? Can I just stop you? That CapEx and we've talked about this many times on the show. What is that being spent on specifically? And I know that one of them is new clouds, but is it just is this just more data centers? Like is this just we're spending this to make data centers only? Like is that what it is? >> Yeah, more data centers, chips, a lot of it is going to go through Nvidia. We talked about Micron, right? Memory. This is this is a lot of it is going to go into memory as well as memory prices are, you know, essentially doubling every year. And yeah, a lot of these chips, data centers memory and a lot of the downstream and upstream companies are going to tremendous like supply chain companies are going to benefit tremendously from all this CapEx spend. >> And that money is coming from where? Like are they raising it? Is are they just taking their profit, their liquid cash to invest that? How does that work? >> Yeah, they're essentially taking their free cash flow from all the major major companies are like Google, Amazon, Microsoft, and Meta. They're essentially spending all their free cash flow and as well as Google. We saw Google going to the bond market to raise billions of dollars. They did a 80 billion dollars in private or equity raise. Meta is raising billions of dollars in the bond market. SpaceX had their big IPO, so they raised about 80 billion dollars through their IPO. So these they're essentially taking all these money that they're you know, from their company and spending it on the supply chain and the downstream and upstream companies. >> And the hyperscalers are just Google, Amazon, Microsoft, and Meta and now SpaceX as well is part >> Correct. Correct. I would also include Oracle in here cuz they spend a lot of money as well, but they're not even included in here. So >> Okay. And that growth projection is Okay, so right now you gave us the 2024 number. This year it's it's it's around just under 800 billion and then we're looking at 1.2 trillion, 1.4 trillion in each of the next respective year. So, that spending is just going to accelerate. It's going to basically double within 2 years. >> Correct. Exactly. >> Okay. Okay. Okay, continue cuz I I don't even know what Neoclouds are. So, so tell me more. >> Yeah, so So, this is actually before I explain Neoclouds, I want to show you the the revenue forecast. Um so, in 2023, if you look at the chart, um revenue was basically nothing. That Neoclouds made made nothing. But by 2025, across 25 billion, um and Synergy Research is now projecting a 58% compound annual growth through 2031 with total revenue approaching about 400 billion by the end of that window. Um so, essentially they're going from zero to 400 billion in like a few years, like 5, 6 years, which is insane to think about. And that is the tide and the question uh and and the question I want to answer today is, given that tide, which boats actually want to be sitting in because not all of them are going to survive. In fact, majority will fail because of the enormous leverage underneath these businesses. Um so, this is actually a good segue to explain what a Neocloud is. Um because I think uh people overcomplicate it. I don't think it's that complicated. Um but essentially, a Neocloud is uh Neocloud is simple. You buy Nvidia GPUs, you put them in data centers, and you rent them out to AI companies that need compute. That's essentially the whole model. Um they also secure power supply contract, and they basically do everything to get a GPU running inside their data centers. So, you want to think of it like AWS or Google Cloud, but stripped down to basically uh one thing. It's basically giving you raw uh GPU access for AI and machine learning workloads. Because if you if you could put up the next chart, LG, it's from Summit Analysis, is look at this chart because of this shows you how crowded this space has gotten. On the far left, you have your traditional hyperscalers. Um you have Microsoft Azure Oracle AWS Google Cloud. These are the giants that have been doing clouds for decades. Then you have the neo clouds, which is you have your CoreWeave, you have your Lambda, you have your Crusoe, and you have your Nebius. These are the ones that have the balance sheet, the customer relationships, and the power and serve infrastructure to compete it at the highest level. Then if you look at the middle, the there are a lot of emerging neo clouds because a lot of the Bitcoin miners essentially converted their, you know, business models into neo cloud because crypto is trash right now. Uh so you have Together AI, you have Applied Digital, um you have RunPod, uh you have GMI Cloud, and the list essentially goes on forever. And this is a bit of a older chart as well, so there I'm sure these this is well over 100 neo clouds emerging neo clouds right now. Then you have your um sovereign AI clouds as well, which are your government-backed infrastructure plays in countries like India, Vietnam, Japan, um you name it. So what does this chart all mean? So basically, what this means is market is about to go through a brutal consolidation phase. You cannot have 78 emerging clouds all surviving in a world where economics only work at scale with cheap finance financing and investment-grade customers. So most of the middle section of this chart, uh like the ones I named, will eventually get wiped out or get acquired or I'm just, you know, run out of money and get shut down. >> Let me just rewind just so I understand. So, Neo Cloud is basically a company that's taking that's that's creating compute that is only for AI. >> Correct. >> That's That's basically it. So, there's like all these other companies that create compute, they also create it for a lot of other right? Like all the other everything else that needs to run. But basically, CoreWeave and Nebius and all the other names that we know, they have gone out and created And do they have their own data centers or they're using space in other data centers? Or how does that work? >> Yeah, they have their own data centers. >> They have their own data centers and they're like, "This is a data center just for AI. Come rent from us." Like basically, rent some compute and they Okay. Okay. So, that So, if you are Who are their customers? >> Uh Meta, you have Microsoft. These hyperscalers essentially >> Okay. So, the hyperscalers are their customers and they'll come and be like, "Listen, you have all We need some extra uh data centers for our AI. You have a data center that is just for that and I'm going to rent from you." And as a result, CoreWeave and Nebius popping off or or or all these other companies are designed to do that, right? >> Yeah, exactly. >> Okay. >> Because this is This is actually perfect cuz like Yeah, you're like the difference between a Neo Cloud and hyperscalers that that actually matters a lot because like you said, LG, like AWS, Azure, Google Cloud, these are massive massive platforms offering different services like your databases, you have your networking tools, you have your products, you have your storage, analytics. The list goes on forever. Neo Cloud's basically does none of that. It basically does one thing. It gives you the most powerful GPUs available. Um and you can run your workloads on that on that GPU without the extra software layer slowing you down. And it's also very super fast networking built specifically for these massive AI runs. >> Here's my general question for this I don't understand about even the emerging Neo Clouds, right? Cuz something like you'll I feel like we're going to talk about CoreWeave and Nebius a little bit more in a bit. Those are companies that have been around for a while. Some of these emerging ones, are these new companies? And if so, like they're not new companies. Okay, so they they already exist. >> Some are new, a lot of them are not. They're just a lot of them are just Bitcoin miners that are converting their assets into Neo clouds. >> Got it. Okay, so like Iron or however you want to pronounce it. Yeah, they so they is Iron is a Neo cloud company. >> Yep. >> Right. Okay, so they were like we have these we have these massive centers for mining Bitcoin. We're going to be done with that. We're going to convert everything. I guess probably change some of the chips, change a bit of the infrastructure to modernize it and off we go. Now you can rent our Neo cloud, right? You can use our facilities for your AI capabilities. Got it. Why wouldn't the hyperscalers just buy all these companies or just buy their the facilities? Like if I'm if I'm starting a new Neo cloud company, how how am I how am I building that in a way that Meta or Microsoft can't? >> Well, they are. The problem is the hyperscalers or the Neo clouds companies got a really had start over, you know, all these hyperscalers cuz all the cap backs and everything didn't come in till, you know, earlier last year 2025 I would say, you know. So they really started investing last year. But these Neo clouds saw this coming, you know, companies like Nebius, they were building from 2023 onwards. So they have they have a head start. And also what's interesting is they already have like these power supplied already locked down. Whereas, you know, these hyperscalers if they were to build it, they still need to go get supplied the power supply that they need that the market is like you essentially have none right now. Like it's so hard to get power. And that's why you see, you know, companies like Bloom Energy like run up massively these days because of the demand for onsite power. So >> Got it. Okay, so then that leads to another bottleneck, which is power, which is energy because everybody's trying to do this. Got it. Okay. Okay. Okay. Cool. Okay, so that so this is this is a it's a big pie and there are a lot of competitors at all scales basically in the neo cloud space, right? Okay. >> Correct. Correct. But with that being said, I think the business model underneath uh is actually really interesting and understanding that will tell you will you know, tell you what actually the risk are. Um think about it like this. Um the moment a GPU ships, right? It's losing its value. Every single day it sits in a warehouse unsold or installed but not rented out yet is depreciating and you basically earn nothing. Um so the entire game for these neo clouds is one thing. How fast can you get I bought the GPU to to whether someone is actually paying for it. And then on top of that, there's also these financing problems that are that are happening, too. Nobody is buying billions of, you know, billions of dollars worth of GPUs on their checking account. You're essentially borrowing to do this. So you also have the depreciation cost and now the interest and the rent you collect has to beat both of them to see an actual dollar profit. >> Got it. Okay. >> But yeah, so this is a hard this is a hard business to make, you know, money in because you have all these fixed costs and 1 gigawatt nowadays costs $100 million to build. So this is this is a lot of lot of money going into this to build out these gigawatts of power uh of uh compute capacity. Um but to be completely fair, the the way neo clouds, you know, solved the financing problem is actually kind of a smart because essentially you go to a customer like your meta or Microsoft and you get them to sign what is called a take or pay contract. That means they're on their hook whether you whether you actually use the GPUs or not. It's like a commercial lease, right? You sign it, you pay it. The landlord doesn't care if you like the business is slow this month. Then you take that signed contract and you go to a bank and you borrow against it. So, the bank looks at that guaranteed cash flow in the same way like the they look at it like a toll road, you know, it's predictable, low risk, and they they can lend it cheaply. Um so, that financing is what essentially made this whole industry possible and and it is scaling at a massive massive um uh scale right now. Um and without that, none of this works. So, yeah. >> Okay. How much of the um I I think you covered this at the start, but just remind me how much of that CAPEX is expected to be Neo clouds? Is there a figure for that? >> It It's hard to figure out. There's no exact figures. We can make these estimates, but um yeah, it's really hard to figure out, but what I can tell you is compute we're severely compute constrained in in in the world right now. Um and Neo clouds will benefit greatly from that because they're they're actually the ones who are taking electrons and converting them into tokens, you know, they're the only ones who are who can really do that. Sure, hyperscalers can't do that, but they have the power, they have the GPUs, they have built these amazing amazing um you know, a relationship with Jensen. So, yeah, they're they will benefit from this massively. >> Got it. Um I think one the next thing we want to talk about too, Melvin, was what happened recently. I alluded to the dip um and this was largely because of some news from Meta. >> Correct. Yeah. Um so, Meta Actually, before I get there, I I think I want to talk about what Jensen's kind of doing with um with these Neo clouds because that'll actually set the tone for Meta Meta news that came out. Um so, we all know how much like Jensen loves Neo clouds, and he has put numerous $2 billion investments in several of these like CoreWeave and uh Nebius. Um and Jensen Huang is not just selling GPUs to neo clouds and just, you know, basically collecting a check. He's essentially financing the construct financing the construction of an alternative AI infrastructure layer designed to completely uh compete with the hyperscalers. Think about what Jensen has done here in the last like 6 months. He has put 2 billion in CoreWeave, 2 billion in Nebius, equity in Lambda, N Scale, Crusoe, and Nvidia is starting to even backstop GPU financing directly, uh meaning lenders can now fund geo GPU clusters more easily because Nvidia is essentially guaranteeing a minimum revenue for um revenue floor on the underlying hardware. So, what this means is that Nvidia is telling banks, "Lend these lend to these neo clouds to buy our GPUs and we'll make sure the revenue is there to back it up." And that solves the single biggest bottleneck for names like Nebius and CoreWeave cuz how do you finance 16 to 20 billion dollar in CapEx this year without either torching your balance sheet or like diluting your, you know, shareholders? So, if Nvidia essentially becomes backed financing, that becomes the standard for credible neo clouds and they can raise cheaper, scale faster, sign more, you know, flexible customers. Um contracts and it's just a snowball effect from here on out cuz what they're doing essentially is building up uh infrastructure now and later they can reap the benefits of all this like the profits and stuff. That comes later on in in the in the stack. So. >> Real world assets like funds, treasuries, and private credit are still running on rails built decades ago. Gated, paperwork heavy, slow to settle. Everyone's talking about tokenizing them, but far fewer can actually do it and do it without cutting regulatory corners. Securitize can. It's the SEC regulated infrastructure bringing real-world assets on chain. Nine years in, native tokenization not wrapped, backed by BlackRock, Morgan Stanley, and Cathie Wood's Ark Invest, and chosen by the New York Stock Exchange, VanEck, BNY, and Apollo to do it at scale. It's the regulated bridge between traditional finance and crypto. Tokenize the world at milkroad.com/securitize. >> Everyone's tokenizing stocks these days, but almost nobody's doing it right. Thin liquidity, prices that drift from the real thing, dividends that just vanish. Bitget Stocks 2.0 is different. Real Nasdaq and New York Stock Exchange depth through licensed brokers. Prices mapped one-to-one, dividends paid to your account in real time. Plus, you get the lowest fees in the market at just .04%. And you can trade them like any other crypto as margin, in earn, in grid trading. Tokenized stocks, finally done right. Head to milkroad.com/bitget to get started. >> Okay, so so hold on, let me just recap that because it was a little complicated for me. So so Jensen Huang is is from Nvidia is basically taking stakes in the neo cloud companies to combat the hyper scalers? >> Yes, correct. >> basically have a hyper scaler in his back pocket through these investments? >> Exactly because because you know, if you look at the last in the landscape the last couple months, everybody's building their own custom chips. You have Microsoft building their own. You have Open AI that launched jalapeno with Broadcom. You have Anthropic with their Samsung chips, rumors of them building with Samsung. Deep Seek is you know, designing inference chips. The hyper scalers are essentially trying to escape the Nvidia landlord, right? So Jensen, in order for Jensen to fight with that is fund these neo clouds, put all the GPUs, like his own GPUs, into the into their data centers, and that's a guaranteed revenue for them, and Nvidia's ecosystem is going to continue to grow, and then the the rest of the ecosystem will continue to grow cuz he does not want an ecosystem where only the hyperscalers thrives. He want a full ecosystem where neo clouds, the the the chip manufacturers, everyone thrives cuz he will make more money that way. >> So. >> Got it. Okay. Okay. Okay. So, he basically he doesn't want there to be this off balance of winning companies. He wants everybody who's using his chips to succeed, which will which will which will further benefit him. >> Exactly. >> Got it. Okay. Okay. Okay. Got it. So, let's go back to Meta cuz we we said we'd come back to that. So, what what did Meta do recently? >> So, Meta So, there was news that came out that about week and a half ago that essentially saying that Meta has built so much capacity that they do not need um to be buying GPUs from third parties. Uh so, now this this is all false cuz essentially what that news, you know, is saying that the demand for these neo clouds is going to dry up. The neo clouds will suffer because Meta has actually signed deals with Coreweave Nebius for about 50 to 60 billion dollars in backlog, but this is a completely stupid and idiotic take, to be honest, and the data that came out last week proves it. Because uh shortly after a week later, Meta announced that it's expanding its Richmond Parish, Louisiana, data centers to 5 gigawatts of compute capacity. Um that lifts the announce the investment from 10 billion to 50 billion dollars. This came out actually today. Um they also announced a $10 billion investment to build their first data center in Canada for 1 gigawatt capacity. And so this is not a behavior of a company pulling back from on computer, but they they realize they that they need more compute than ever before. And Zuck is just going to keep on building because they know they need it. Cuz if you just pull up that chart on the meta, yep, meta's um compute capacity. So basically meta's compute capacity goes from 7.5 gigawatts at the end of 2025 to 12.3 in 2026 and 17.8 in gigawatts in 2027 and 21.2 gigawatts by the end of 2028. Even with all these compute going online, there is they're investing more money than ever before because the demand for this is essentially exponential at this point. And there is no signs whatsoever of this stopping or slowing down anytime soon. Now, here's why I don't think meta will be selling their, you know, excess compute capacity because they will essentially have no compute to sell because meta's super intelligence continue to push their frontier model training. They're, you know, we saw their spark, their new model that came out last week. So they will be using a lot of their compute capacity for that. They also planned roughly 10x scaling of meta's ad recommendation systems, which requires enormous amount of inference compute. And a llama model, I know they're, you know, they kind of sucks, but they're still deploying the open source llama models in every meta product. So meta's compute compute capacity is only ever going up. And Zuck is even talking about, you know, wanting a wanting to put more compute or more um compute in the glasses, you know, which is going to need tremendous amount of inference. Um so, Meta is is not going to is not going to sell their capacity because they're not going to have any. So, the narrative that took down, you know, neo clouds was Meta has excess uh compute capacity, the capex is speaking, but that that is not like that is not what's happening here. Meta's capex is only accelerating. Their compute is scaling 9x, and they're essentially they're potentially becoming a compute reseller on top of their demand, but they're not going to do it. There's no way they can afford to. So. >> Got it. Okay. And then then when this news came out or this kind of like uh false news or misleading news, it caused a dip. >> Correct. >> Across all neo clouds, right? Okay. >> Yes, correct. >> Right. >> So, that that's >> Cuz the market misunderstood it, right? Basically, the market misunderstood what that meant. It was a little over reactive. >> Correct. And so, that that's actually a good segment to where the opportunity is cuz there's a lot of names that are down that I'm I've basically added to my positions or um so, first one I want to talk about is CoreWeave. Um so, their origin story is one of the better pivots in recent business history. They were founded in 2027. They were Ethereum mining operation. Um and then they pivoted to GPU as a service in 2019. Um so, they were early. By 2023, they were renting H100s at $8 per hour during the peak like GPU shortage. They basically saw this trade before anyone else did, and they positioned it very accordingly. Um today, they are the operational scale leader in in the in the space in the entire neo cloud space. And if you look at the chart, uh I want to everyone to see what the customer commitment picture looks like um just this year alone. Um Open AI committed $11.9 billion through 2030, another 4 billion through um 2029, another 6.5 billion to 2031. That's over 21 billion just from Open AI. Then you have Nvidia signed a 6 and 1/2 billion 6.3 billion deal with um with through them 2032. Nvidia is literally a paying customer. It is also an equity investor as well. Um then you have Meta. Meta is going to committed 14.2 billion through 2031 and then added another 21 billion um through 2032 in March of this year. Anthropic also signed a multi-billion dealer deal with them. And then you have Jane Street, one of the one of the quad quantitative firms, trading firms, um also signed a $6 billion deal in April. Um so total uh deals from that uh from that table alone is roughly six 6.3 63.9 billion. And that is not even the full backlog. That is just from this year. Their full RPO is 99.4 billion dollars. >> What's RPO? >> Uh backlog. >> Backlog. Oh so so what is it what is it what is backlog mean? >> Yeah, remaining performance obligation which basically means they have yeah, they basically have um essentially like 99 billion dollars that's supposed to come in in the next, you know, X amount of years. >> Holy Okay. >> Yeah. But what's interesting is they essentially have a bigger backlog than their entire market cap. Their market cap is around like 50 billion dollars. And they have a yeah, and they have a backlog of, you know, 99.4 billion dollars. And and as forward demand for compute increases Coreweave benefit from this greatly. Coreweave currently has 1 gigawatts of live active data data center capacity and they also hold 3.5 gigawatts of contracted power capacity across this portfolio. So this is what I mean when I say like they have these deals already signed. They have you know 3.5 gigawatts of power. And the company is actually looking to scale about 8 gigawatts of active power by 2030. So they're expanding massively. But to be fair there is a lot of problems with with this company. Not a whole lot of problems. The main problem is their debt. They have a total of 24.9 billion dollars in total debt. That they have roughly interest expense of 536 million in a single quarter. Net loss of 740 million dollars. But what I do want to what I do want to say here is the the model their business model is designed to be front loaded. Free cash flow for them doesn't even turn positive till 2029 2030. As revenue grow faster than the debt debt stack. Like I know they have a lot of in interest but that should that'll be backed backstopped by their revenue growth and by the time they get to 2029 and 2030. And they're essentially just building the infrastructure right now and all the revenue will come in later. Same this is the same argument you can make for the hyperscalers right now and kind kind of the reason why they're kind of down is because they want the you know the revenue to come in. But you know it won't come in for another few more years. That's the same deal with Coreweave. >> Right. So similar to everybody else they have all these deals signed for a long time and those are commitments like we just saw on that slide right? It's like that's a lot of great money but it's not it's not 63.9 billion today. It's not money that's delivered immediately. It's CapEx money, right? Like we said at the start, it's part of that 1.2 trillion that's going to come out and on paper that's what's been committed to, but those are timelines even we're looking at right now is like four four to six years in there. So that they will get that money over time, >> Correct. >> but it's not money in the pocket right now. >> That is correct. >> And same with a lot of people, same with a lot of these commitments over the next five years and 10 years and however long it is. >> Absolutely. And but one of the reasons or one of the things is companies are continuing to raise their CapEx. So all that money will eventually get, you know, paid out. So I expect these CapEx to continue to rise into 2028, 2029 even. Somewhere around maybe 2029 it'll probably peak, but but by that time hopefully our infrastructure layer is all built out. Then 2029 probably is when Elon will probably launch and start to build out the space infrastructure needed for in space. So that'll be that'll be interesting to watch as well. >> Tell me about another company that you like in space. >> Yeah, I think my favorite one is called Nevius. It's at around $211. They have a rough market cap of $52 billion and like I said this is one of my favorite stock in this entire sector. This company has a really cool backstory as well. They were actually a successor to a company called Yandex. If you don't know what Yandex is, it's basically you can think of Google of Russia. They had search engine, they had their search engine, maps, ride sharing, e-commerce, cloud businesses, and essentially Yandex was the dominant internet platform across Russia for several decades actually. Then what happened was then in Russia Russia actually invaded Ukraine in 2022 and that basically changed everything. Uh because Yandex ended up you know selling their entire Russian business for about four to five to six billion dollars at you know at fire sale prices. They basically handed over a multi-decade you know empire for cents on the dollar cuz they just wanted to get out. Um the company essentially renamed themselves to Nebius and started rebuilding from scratch as a AI cloud business using and using that proceeds to build out um from that build out the AI infrastructure. Um and they have you know when they left Yandex they had a whole bunch of engineers. They had a killer team. Um so that's one of the reasons why I like it because they already built companies. They were in business for decades. This is a great company. This is a great the founder is incredible. If you listen some of his interviews he's an impeccable impeccable um founder. Um now let's get into the numbers because they're accelerating at a speed that is like very hard hard to fathom right now. Um Nebius AI cloud essentially grew 841% year-over-year in quarter one of 2026. They did 389.7 million in AI cloud revenue in a single quarter growing from zero to just you know that number in two years. Um their total revenue came in at 399 million at 74% growth margin. And their full year guidance for this year is three to four billion um with or in revenue with an annualized rate of 79 billion dollars by end of the year. Um to put that in perspective it took AWS um you know Amazon Web Services much much longer to reach that revenue at scale um and Nebius just basically did that in two years. Um and they also have the very long backlog as well like CoreWeave. They signed 17.4 billion dollar deal with Microsoft in September of last year. That was Meta also committed 12 billion dollars in capacity for 2026 and about 27 billion dollars worth in total. And then top of that, like I said, Nvidia put in 2 billion dollars of investment in March of this year. That means, you know, they get, you know, the cool thing about every time Nvidia invests in one of these companies is that investment means they get GPU allocation faster, deeper technical integration, the the access to the first generation next generation hardware. So, actually the funniest thing was the the Nebius actually posted something today that they have taken a delivery of Nvidia Spectrum 6 1002.4T Ethereum switch from Nvidia today. And they got early access for that because they they their their deep relationship with Jensen and you know, Nvidia. So, this is why I love companies that invest Nvidia invest in cuz Nvidia is essentially the kingmaker of this AI infrastructure, you know? So, the more you can get closer to Jensen, the better it is for your companies and how your company is going to grow. >> So, is that is that is that is that one of the the reasons you're you're like is that the reason you're most bullish really is the affiliations Nvidia and Jensen taking those positions or is it just the general CapEx need and and commitments already? >> Yeah, affiliation, CapEx, the need for compute, and also because Nebius has a killer team. I've, you know, CEO is great, their entire team is full of very very experienced engineers running Yandex and they have other killer businesses as well which I will get into in a little bit but and also what's cool is Nebius has contracted over 3 billion 3 gigawatts of power capacity globally globally and is targeting about 4 gigawatts in total by this year. What's cool is like there is a need for massive amounts of compute just not not just in the US outside of the US like in Europe you know so Nebius is the leading leading for that so they broke their you know what's cool is they broke their first US scale campus campus in Missouri with meta this year then they announced a second one in Pennsylvania in Europe they're actually building a 310 megawatt data center in Finland they have three locations across UK they have deals with like the British government. So this is genuinely a global like build out that's happening all across the world and Nebius is essentially you know building that out but what's what's different why I like this better than Coreweave is because Nebius actually has 9.3 billion dollars in cash on their balance sheet without any meaningful debt net debt versus like you you look at Coreweave which has a 25 billion dollars in debt debt. So if something was to go wrong like or let's say the rates goes up or something you know obviously as the rates goes up you're going to have to pay more in interest they're essentially bulletproof at the moment they they have they have money on the balance sheet so that makes them a way better company than Coreweave. >> How guaranteed is that money that's committed to Coreweave and Nebius? >> A lot of it is it's the agreements that I talked about so even if they even if they like they have to pay up like there's no there's no way around it. So, um a lot of it a lot of it is already baked in. So, it's very very bulletproof. >> Mhm. Mhm. Okay. Uh you mentioned other businesses that Nebius has. Is that Is that Is that We're almost out of time, Melvin. So, is that a different podcast or is that a today thing? >> No, I can I can briefly get into that real quick. Uh one of the one of the companies that they own a stake is in is called ClickHouse, which is basically a high-performance database company where Nebius holds a significant portion of the stake. And based on the recent fundings, they are worth about $15 billion. Their stake is worth $15 billion. So, that is roughly 30% of and and you know, or 15% or 30 Yeah, 30% of Nebius's entire market cap. Um so, and ClickHouse is essentially supposed to go public sometime in the next couple years. So, they will you know, Nebius will get massively benefit from that. Beyond ClickHouse, they also have AveRide, which is autonomous vehicle technology. They have Toloka in AI training data and data labeling Triple 10. There's a lot a lot of different companies that Nebius is part of. Um so, because you have to understand that this is a well-established company back in back in Russia. So, they have they still have stuff left over from there. And yeah, this is one of the reasons why I remain extremely extremely bullish on Nebius. >> And Melvin, you've been adding to both positions during the step, correct? >> Correct. I've been I've been buying heavily. I think I think right now um market's giving you a great opportunity to buy these um stocks. I know they're a little bit beat down. I think Nebius ran to about 300 bucks this year. Now, it's around 212. I expect fully to expect them to finish by 300. I can even see them go into 325 by the end of the year. Um Core Weave, on the other hand, I think they're going to 150 by the end of the year, which is I I know it's essentially going to double in price, but I think the market will eventually find out that you need these uh Neoclouds and compute capacities heavily constrained, and these two will benefit massively from that. >> Awesome. And for anybody wondering, Melvin first called Nebius it was like the first day of portfolios, man. This was like when we first launched the portfolio uh product in Milkroad back in February. Uh Nebius was like your first buy and it was at 95 bucks. So it's been an excellent call so far and and you're calling for higher uh on both companies and I think I think you've given some solid points today, man. I think that makes a lot of sense for for a lot of good bull cases for these and you know, the old adage buy when others are fearful or just buy when Melvin buys, too. You know. >> And I also have a couple couple ones in my uh you know, Neoclouds that I'm watching right now. So I think I will just probably share that with Milkroad members later sometime this week and you know, they can benefit from that greatly. >> Oh, somewhere some of them some of the ones on this giant list that you showed us? >> Possibly. Possibly. This is a sneak >> Who knows which ones it'll be. Uh we'll save that we'll save that for the pro members. Melvin, uh excellent chat, man. This is this is I am informed now. Now I understand and now I know who the players are and and what the need is and what the next 5 to 10 years look like. So I appreciate it, man. Thank you for all the info. >> Awesome. Thanks for having me, LG. >> Want to stay ahead of the biggest technological shift in history? Subscribe now to get insights straight from the sharpest minds in tech and finance. Quickly you'll note this show's for educational purposes only. Nothing here is financial advice. Investing always carries risk. Never invest more than you can afford to lose. Thanks for tuning in. See you in the next one.