Nvidia Bet Big on These 2 AI Stocks — Which One Wins After Earnings

Nvidia invested in two AI Neo cloud players and they both already reported earnings. Yesterday, Neil and I actually took the time to do a very, very deep dive on one of their earnings. Now, this morning, another one reported earnings and I'm not going to lie, I'm not a hater, but there were a few things that I got a bone to pick. So, hopefully Neil can answer some of the questions that I have here and at the end, we'll decide which one of us or which one we pick to be our favorite investment long time. So, let's take a closer look in today's episode. Now, before we continue with today's episode, if you want market beating stock picks from our analyst, make sure to check out the pin comment and the description. Using that link gets you a promotional offer as our thanks for being a viewer. Thank you, and let's get back to today's episode. Hey Neil, welcome back to another episode. Uh, today we're going to take a closer look at another NeoCloud player. Yesterday, we looked at Coree. They just reported earnings and um earlier this morning we had Nebas uh report earnings. These are two players that Nvidia has invested uh money in. Uh these are two players that are investing a lot of money in buying Nvidia GPUs and are really important players in my opinion in this AI ecosystem. Uh I would say a lot of retail investors know at least one of these two names um just because they're very popular and the overall market is pretty excited about the demand for AI uh and AI inference. So Neil, I know you cover Nebas a lot more than me and you understand this one a little bit more. Can you give us kind of some of the top numbers for Nebas before before I go out here and kind of give you the three complaints that I have? First of all, I I'll just say I am surprised that Nebus is up more than corre at the time of recording this video. I believe Nebus is up almost 30% right now while Corife is up 20%. And remember, Nebus was a name that was up double year to date while Corore was struggling to stay flat. And I mean, we've covered it yesterday. The numbers were were excellent for Corore and the thesis became stronger. So I I I just find it weird that the market is still skeptical on one name, but we'll discuss all of that later. For Nebus itself, revenue-wise, they're of course growing extremely rapidly because they come from a very very small base. So revenue was up 454% year uh yearover-year. AI cloud ARR was up almost 600% to $3 billion. AI cloud adjusted EIA margins at 50%. And now the year end contracted power target was raised to 5 GW. Previously that was over 4 GW. Now all of that is great but to me what was much better was their comments on pricing which is something that Corif discussed yesterday as well. How fast they think they could be moving from here on out. So they first of all for the quarter they closed four landmark AI deals. The average TCV which is a total contract value was over a billion dollars. 20 to $25 million ACV per megawatt. So the contracts worked for Reflection Coher an unnamed Neolab US one we don't know which one. And a large US quant trading firm again we don't know uh the name of that either. As for their asset light business, which they announced, I believe a month or so ago, that should unlock 2027 capacity and much more. They do plan to add 1 GW per year starting in 2027. So that would allow them to move much much quicker. We can go over that asset light model in a bit. What was new this quarter was one the pricing what is going up. So they went from around $12 million to now 20 to $25 million in Q2. And they already seeing just like we've said with Core Reef uh yesterday, Cororewe did take a small chunk of of their capacity and they're now looking at more short-term contracts because there is a premium there. Nebu said that those short-term deals they are now looking at 40 to $50 million ACV to me which is a complete gamecher. They also told us that they could sell out 2027 capacity right now if they wanted to, but they are not doing it because they can do these short-term deals and capture a huge premium. Now, of course, you can be skeptical and say we don't believe this, but based on the numbers, I highly doubt that they they are just making things up. And so, to me, this was again a a very very good quarter by the company. the fact that things are moving this quickly. Capex wise for for the full year stays the same between 20 to $25 billion. They did say that connected power is still on track to reach between 800 megawatt to 1 gawatt by year end. And from my understanding, it's going to be active power by mid 2027. So around 800 megawatt to 1 ghaw mid 2027 of active power which we could we could start to calculate what the AR could be. But if if you take the 20 to $25 million or some premium from the $40 to $50 million you can start to maybe uh calculate what the potential AR could be along the way in in 2027. This here is basically the asset light model. On the one hand, you have the partners. They provide the land, the facility, the power, the hardware, capital, and the operations. And Nebus comes in with the full stack AI cloud software system architecture and reference designs. Global go to market and demand. And of course, the high margin service layer is all for Nebus to capture. As for Violent, Vinand is part of the capacity. Vinity in New Jersey. It's uh it's part of the capacity for 2026, not for 2027. And they still believe that the project will be on time despite all of the headlines and maybe saw some legal issues there. They did switch to Bloom Energy. So that's the energy fuel cell uh power solution which should let's say reduce the complications against against that data center there. So, so far they believe Vinlet is going to be on time for uh this year. As for 2027, and then I leave the the rest to Jose for 2027, the guidance for now is still the same from what we've got previously, but they're going to give us fiscal year 27 guidance either later this year or probably in Q4. But for now, it's between 7 to9 billion. Group revenue 3 to 3.4. 4 billion in fiscal year 26 40 40% approximately adjusted Ebida margin and the capex as I said previously was 20 to 25 uh billion uh dollars now to me we are going to see a huge acceleration in growth in 2027 because I I I do think more capacity is going to come online towards the back half of this year we might get a deal but then again that's that's speculation from my part know that Google is going to look add some new cloud players, right? In Q3, they told us that. So, Nebus could be one. Again, it's just pure speculation right now, but if you look at the comments from Google, if you look at the comments from Nebus, maybe maybe there might be a connection uh there. But I'll leave it to Jose right now and then we'll we'll get back to some uh some other numbers. Yeah, Neil, I mean, I thought it was a great quarter uh for Nebus and um I I think even what can we take out from here just for the whole ecosystem or just to kind of confirm of what kind of core we've reported yesterday. Some of the things that I really liked was they continue to tell us that demand is not the issue. There's so much demand for for the AI infrastructure right now. Uh the other thing that they're telling us is uh pricing is going up on older GPUs and newer GPUs. Uh right. Uh so so that's something that we heard as well. So it's great to kind of get these validations um from another party. Now Neil, like I mentioned, I have three complaints about Nebus and it's also for the fullest. I know Core Weef and Nebia sometimes get put in the same bucket cuz they're Neoclouds, but they're actually both running different businesses and it's not that one is better than the other, but it's good to understand that they are different. For example, uh Cororeweave, a lot of their contracts tend to be five plus years, right? Okay. And and when you have that ability, you do have the ability to get this massive financing and kind of get debt up front to build these massive AI infrastructures. Nebas on the other hand is really focusing, they mentioned their core revenue is coming from like that two to threeyear contract. Um uh and and those two to three years tend to get a little bit better margins than five years. And just like Neil was mentioning, like now they're starting to enter the shorter term contract um which could be like 3 to 6 months. and those tend to have even better margins than that two to three years which tend to have that better. So um we we're seeing other players also shift into shorter com um contracts as well. U but that's one of the main difference between these two players and then Nebius is really focusing more on the enterprise players where they can really upsell a lot of more software stack solutions and that also helps out with the margins. Now that we understand the diff at least some of the top level there's other differences uh between the two three things Neil that I hated. First thing is active power. Why can't they just give me a number of active power? And for those that are not familiar active power is the power of GPUs that are running and actually collecting revenue. They when they they even gave us right that revenue that they can collect per megawatt. So, if you're giving us that information, right, that information, if you gave us that information plus active power, that's a a wealth of information for investors. Um, I I know if you read between the lines, Neil, I think you you broke it down that they said 800 800 megawatts to 1 gawatt of contracted power and that by the first half of 2027, that start collecting revenue. So I think like you mentioned you based on that reading you can assume that you'll have that 800 to 1 gawatt of active power the latest the end of of the first half of 2027. So I I like that that thinking there. Um but I I I wish they would have actually given a little bit more information on that. The second thing, Neil, that I was not a fan of, and I think this is great for the business, but if I was a customer, I wouldn't like it. And I think when you're running a business, you have to make sure you're you're you're you are in a fine line of doing good for the business, but also doing good for your customers. >> Before you say it, can I guess >> is it going to be the auction one? >> It's the auction and then the other thing, another thing that you already mentioned. Um but let's start with the one you mentioned first is they openly said we're just holding capacity that we don't want to sell out just because a few quarters from now we can probably get better pricing. Like if I was a customer and I'm like you're not it would get me mad if I was a customer. It's like you're telling me you're sold out but you just went on an earnings call and said you're not sold out and you're just kind of like shaking me off trying to take all my money. Um, so it's like I it it just left me the rub way, but I agree that it makes the most business sense, right? It it's more of like if I was a customer, how would I see it? And the other one, just like you mentioned, the auction thing, right? If I was to get an auction is telling me you don't know what the price of this market is right now, and you're just it's whoever has the most money. Again, it's great for a business, but if I was to come and rent compute from you, just give me the order. I'll place the order if I like the price. Don't let me have a bidding war, right? Don't let me have a bidding war with the next person. Uh go in and out. So, um both those things, Neil, I think are very good for the business and it's going to help increase margins. It's more of just like it leaves a bad taste in my mouth. I wonder if customers right now it's like there's a supply constraint demand. I'll just have to take it. But, uh do they remember things like that later on? And and is that going to leave like a negative term? Uh I'm probably nitpicking here but uh it's just few things that really caught my attention. >> First of all the auction part again first of all like you said it's very good for the business which means it's very good for for investors. Then as for the well the auction part and and the other part actually goes goes hand in hand. So as for the auction part they they did it it was a very it was a success for them. So they told us the auction cleared for Nvidia Blackwell chips at the highest price Neuse has ever secured, 15% higher than the highest price Nebus has ever charged previously and 20% higher than their existing pipeline rate for Blackwell capacity. And here's the here here is why I don't mind. Because the winning bidder praised the process, specifically citing the combination of clear pricing and guaranteed access to compute and indicated plans to participate in future auction rounds. To me, this is this is the best the best thing that you you can you can hear as a as an investor of of that company because this means that one they are making more money and two they can charge more because the service clearly is very very good. Otherwise the otherwise the the company that one would not come back and be happy with the with the results. But of course, this is possible because there is a a supply demand disconnect at the moment and and for a long time if we're if we were to believe all of these companies. So, it works. Now, I I do also believe that eventually pricing will figure itself out. Although I do believe that that Mark Zuckerberg also talked about an an auction system when it came to the uh to the token pricing for for the Aentic Works that they that they want to that they want to release just to make sure that everybody can pay the the most reasonable price for for their workload. So to me it it just screams we have pricing power. People like to work with us and we're happy with that. Then going back quickly to to this these are my assumptions. So these are not the numbers from nebus. These are just my assumption and like you said I would have loved to get an active power figure. They like they gave us connected power then hinted that that connected power will turn into active power first half of 2027. So so we can start to assume that these could be the potential ARR numbers. Again not everything comes online at once. So it's a gradual process but these are the numbers that we could be looking at mid 2027 and even if it is later than mid 2027 these are numbers that are way higher than the market expectation right now even if we go between the these two numbers 18 to2 billion is much much higher than any analyst expectation at the moment of course if I'm expecting this to be the number the capex 20 to $25 billion from this year. I I assume it's going to double in 2027. On that note, they also told us that the payback period is now much shorter to 1 year and 10 months instead of two to three years. And I think we've discussed this privately as well and maybe when we covered other companies and also SpaceX talked about this that the payback period is is one year. I think I do believe that the payback period is going to be shorter than one year and 10 months for a nebuse. We're not going to discuss the whole Nvidia partnered with uh Goldman Sachs, Apollo for $500 billion and and all the others and how that could help these types of players because it could cost of capital will come down. All in all to me was was a blockbuster quarter by Nebus and I cannot wait to see what happens in uh in 2027 with Nebus and Core Reef cuz I own both. >> Yeah, I agree. I agree Neil. I mean obviously um I think both are going to do extremely well in this market. I think we are in this complete compute constraint space. I I was really happy uh to me even though I'm not a Nebus investor they provided a lot of information right we didn't get that ACV style information from core at all um so if you are an investor in this space this is where you like even if you don't own it you can grab a lot of information I am curious before kind of we we share our top pick um at the end of of this episode Neil we see that Nebius is starting to do that short-term contract and we had that better ACV 40 to 50 million per megawatt. Curious, obviously with shorter term contracts, the benefit is you get the higher margins. The con is if there is no demand after those three six months, which we don't believe is the case, but let's just imagine is the risk is what happens in six months if you can't re get that that recontract that. Um so so that's something to kind of keep aware of. How put what what are you looking for? Like if you were to have the perfect breakdown of ne of Nebia's core and short-term contract, how would that pie chart look? 50/50, 70 core, 30 uh very top level because I know I'm just throwing this at you. Uh but would you be favored heavyweight at one side or or more of a equal mix? I mean if if they can get such a premium right the the 40 to 50 million which is double the normal well it's it went from 12 to 20 to 25 so but let's say it's now the double the the standard price a a 50/50 split would be great for the for the short term and great for the business I don't think they will they will do it purely because one they do have long-term contracts with with of course a Microsoft and a meta and Second of all, I do think it's just maybe too early, too early to to start doing this 50/50, but I do think it's going to be way more than just 10%. It it would be >> it would be stupid if they didn't, right? Cuz if you can charge such a big premium, why not? Especially this, and we talked about this yesterday with Cory, especially this early in the buildout, why not get as much money as you can so you don't have to go to the market, dilute, raise money, have more debt if customers are now coming in. And they did also say that a lot of prepayments, yeah, customer prepayments featured in roughly 70% of Q2 deals, covering 50 to 60% of associated capex. It's wonderful. >> All right. So, Neil, you said you own both, but if you can only own one, which one would you take? >> The answer here is very to me it's very I I I don't own Nebus. You just own Corore Reef because you only own Core Reef. You used to own Nebuse though, which We feel left behind right now. But maybe that's where all this anger is coming from. >> No, but I I will say this based on the price action, cororewe has a lot of catching up to do. I core the all-time highs for corre happened last year in June. They were at 187. Okay, fine. The IPO, there was a lot of hype. Fine. But it's been a stock that's been underperforming. If you look at the whole neocloud space, it peaked last year in October when when the whole speculative sector peaked as well. But the Nebuse went much higher this year. Corweave has been left behind. I I still think it's purely because of the the fear of debt, right, of of their debt profile. But as we've discussed yesterday, margins are expanding, things are improving, and it's just a matter of time. I think in the short run I I do I can see corore outperforming Nebus just because Nebus has been I mean I think yeah it's been 150% or so year to date. So >> it's uh it's a win-win for me. It's uh we get a lot of questions as well around other Neocloud players and to me the answer is always I just rather go with the two best ones. >> Yeah definitely. I mean I I right now the one I own is Cororeef. Um, and I do believe both of these will do extremely well. Um, to me, the main reason for core reef is I I do enjoy that active power, right? 1.5 gawatt that we just saw in the most recent quarter. Extremely bullish for me. Um, and and just the opportunity to be able to raise to build 500 megawatts in one quarter showcase the opportunity that this company can expand into. Um, and now that they're I I I feel like they're going to be growing at strong levels in active power, but they already kind of passed this threshold of 1 gawatt. And I think from now on, if they continue to add these same levels, you won't get that same margin pressure. Now, Nebas has a great kind of business model and and a great financing. So, I don't think they'll ever feel that same type of margin pressure that Cororeef has felt, but I'm hoping investors are prepared that they don't they do seem like they're going to accelerate that buildout a lot faster in 2027 and beyond. As they mentioned 1 gawatt, there is going to be some compression in margins. Um, and I don't necessarily think it's bearish, but I am going to be curious to see how the market reacts to that change once they start accelerating um in that that investment outpace. So for me it's cororeweave and and mainly because of that active power that they have right now. >> Yeah. And I mean corore they did say what is it the exit rate close to $19 billion probably at the midpoint for for this year. So the number revenue wise continues to be bigger and bigger. So the margin doesn't need to inlect significantly just inflect in a positive way you'll get the dollars needed to then offset that interest uh payment that you need to pay every single quarter which also has increased. I think has been increasing more than double year over year and increases next quarter as well but as they grow more and more that should be offsetting that uh that issue. >> Okay. >> All right Neil I had fun talking to you here. I mean yesterday we talked about Corey. So if you are a Neo cloud investor if you're an investor of AI definitely make sure to check out um the core reef episode yesterday. And hopefully you guys enjoyed the Nebus episode today. And if you guys had to pick one let us know in the comments below. Take care. Have a good day and full