Neoclouds with @EndicottInvests: We discuss CRWV, NBIS, IREN, and what is happening in AI

[00:00:00] Speaker 1: all right everybody welcome back this is a special one for you everybody's been asking about core weave and nebius and i ran so i felt it was time to have another conversation when nate had to bring me on yes what's going on nate what's up appreciate you having me on i'm excited [00:00:19] Speaker 2: dude you know the last time we talked was almost two months ago which is disgusting to say i could double check for fun and i was like oh my it's already been almost two months time flies and [00:00:29] Speaker 1: earnings season is the same yeah oh let's jump right into it a lot's changed i am now a shareholder core weave still a shareholder in uh nebius not a shareholder in iran smart man are they going down [00:00:43] Speaker 2: right now oh man i mean when we talked to i think it was nebius was it like the 220 230s went up to like literally 299 90 or something like they almost reached 300 now we're back down to 195 i mean we touched 181 or something 182 last week thanks to the meta news or that was two weeks ago either way yeah video clouds and the whole it seems like neoclots and semis entirely are also just selling off like crazy which i don't but hey you know it's kind of part of the market [00:01:12] Speaker 1: we did have some big news you started to say meta meta now is expanding the louisiana data center yep 50 billion dollars from i think it was at 30 and 3200 acres the build will not be finished until 2036 [00:01:31] Speaker 2: it's insane i mean so even taking it all the way back to the the new the day of the morning where meta is selling excess compute and for lack of a better term all neoclots took a massive crap and decided the algos decided to sell them off i think nebius down like 18 that day and my literal immediate reaction because i woke up to this news i woke up and i was i honestly just laughed um because it was like what x is compute first of all like my initial my that was my initial reaction was what x is compute um and so and also keep in mind you know meta has a partnership or a contract with core weave and nebius um and so it's like why were they signing these contracts if they have all the x compute um understand a little more i can see kind of you know zuckerberg kind of came on i think he gave the quote of like they're not selling all x's compute but only when they see it best fit essentially so when they see so much demand and they can get such an insane price they'll just sell it because they can make crazy margins on it which is bullish the neoclots sector by the way not bearish uh but of course the ego saw it was like oh my god the ei spend is over it's all doom and gloom capex is over um when reality was literally the exact opposite um and we kind of got that reaction from the market a week later and i met is up a good bit because of it right i think it's up 20 since they kind of talked about that maybe a little less um but it's up a good bit since that news and so i think it's a good angle for for zuckerberg and team and i think it makes sense but it also just shows how expansive this is the fact meta also wants to get into it that means they see value from it um and we're seeing more and more companies want to self-compute so again you know i said this last time but we're just so freaking early and i'm seeing this more and more in my work life is like enterprises are still just trying to figure this thing out like people think that every company is like a google or nvidia where they're just the master of ai engineers no enterprises are slow to adopt slow to figure it out slow to learn but they're willing to spend and build um and we're just so freaking early i don't know what you think [00:03:41] Speaker 1: about that but oh i think that we are still getting ready to go out to dinner we haven't even left the house we haven't gotten a happy hour if you listen to the latest all-in interview with pat gesslinger from ex-ceo of intel he outright said it is not years but decades of the build out that is ahead of us we had larry fink on cnbc today larry fink made the case that we are not spending quick enough or fast enough in compute and the grid from what all these companies are showing us i am not seeing anything slow down i'm seeing a tremendous amount of capital being deployed and i'm extremely excited about core weave and i made a decision initially to invest in nebius because i didn't like core we the debt stack now when i go back and look at the numbers it's a completely different story and core weave is going absolutely parabolic on the revenue and on the ebitda the stock price does not reflect the fundamentals and some people are getting a little uh skeptical about the debt and oh it's down to 75 dollars holy crap it is down to 75 so when i look at the quarterly revenue 111 percent year over year in q1 they went from 982 million to 2.1 billion dollars in revenue this is huge inflection point with the scale up right there and if i layer in ebitda on this and for those of you that are watching the video this is my own system that i built i mean 1 billion of ebitda in a quarter and they're going to continue scaling that now their run rate 4 billion of ebitda likely be more this year and the interest payments are not near that so i'm not necessarily concerned about the debt especially with 99 billion dollars of rpo [00:05:58] Speaker 2: sitting behind the revenue no i think that's a great point and i think also correct me wrong isn't meta's contract take or pay like isn't it yeah like what is it is and i'm an empty shareholder so i'm far more biased nebius but even in this instance like come on it's take your pick like there's no downside for the risk is off for core of in terms of take your pay whereas nebius's deal is not take your pay so i i don't get why corey would sell out even more than nebius like it's corey continues to sell off by the way nebius is kind of it's selling out but it's going a little bit stronger either way i i mean corey have 40 billion dollars is insane to me [00:06:36] Speaker 1: i think the big difference right now is nebius can be looked at as profit taking per se because it's still above 200 it's still up at one whereas core weave is selling off from a level it already [00:06:48] Speaker 2: sold off from true yeah 100 percent um also did you see the the nebius news today i don't know if you saw that i did not so nebius just well i wouldn't i want to call it they create a new announcement revenue stream but they're they are doing a new revenue stream called asset light model so basically what they're doing is making partnerships with companies that are going to pay for everything nebius is going to give them the know-how of how to architect the rack the training for how to optimize the data center and the go-to-market sales motion and they're going to partner and take license fees or a percentage from these partner partnering companies which is interesting um it's a little bit different from the co-location model co-location nebius also own the gpus in the in the network um but nebius announced that today literally and so what that means is they now can scale revenue without having to spend on capex or dilute which is fascinating um and i'm uber bullish on that type of model now that's very i don't know if you've heard of companies like fluid stack um there's a couple others that are very literally a neo cloud where they do that exact quarter sort of a similar thing um but it's really interesting for governments or companies that want their own sovereign data center but don't know they'll have the know-how or the tech expertise to build one but they have the capital and so that was that's what kind of what nebius is offering they're saying hey you know we have the know-how we know how to build we know to optimize how about you come to us we'll help you build it you pay for everything but we're going to take a fee or take a you know a licensing fee or some sort of um fee essentially really and help you build it and optimize it and that's it it's really interesting they actually and they also said in the press release that they have a few agreements already lined up that we don't know how many or who or how much revenues really will bring but i think it's super interesting in the way nebius is thinking about it um you know still a little bit different than core reef who's taking a lot of debt like you said they can pay it off regardless whereas nebius is trying to kind of it seems like push away from that a little bit um and kind of grow as fast as it can without having to dilute and take on debt i mean they're still going to take on debt let's be real but um i just thought this was interesting and what i'm just curious what you thought [00:09:07] Speaker 1: i think it's an interesting aspect i think that we're going to see sovereign data centers become more of a thing because i don't think governments are going to want to keep their data in the cloud i think that we're going to see a lot of on-prem happen running models on-prem i think it's going to be very big i think that's just the next iteration even for hospitals yeah anywhere that has very sensitive data that if it gets hacked you're talking about identity theft etc etc so i can completely see that being a huge line of business i'm just kind of shocked that we're heading into earnings season if we've learned anything from the companies that have currently reported it the party's not stopping the amount that's getting spent on hardware is insane the rpo books are huge we're going to get a new look at the rpo books in a week because you got microsoft meta reporting next week you got tesla reporting next week and then we have alphabet amazon and apple the following week and then we're going to get obviously nvidia and everybody else after that but we're going to see the big ones microsoft and the reason i say that is because they came out and said on q three earnings because they do not report on a calendar year that they are spending 190 billion dollars on capex in 2026 they won't report what their guidance is for the 2027 fiscal year which runs q3 q4 of 2026 and q1 q2 2027. so i'm expecting microsoft to announce 225 to 250 billion of capex because if you're going to do 190 this year and you already said in the announcement with chevron that you are locking a 20-year supply contract with chevron for energy specifically natural gas because you are doubling your data center footprint over the next two years you're not spending less than 200 billion in 2027. [00:11:20] Speaker 2: 100 and i think it's we go through the same sort of whirlwind every three months where people get bearish ai they go oh capex is slowing down then they all come out and say we're going to spend more on capex and everyone goes oh yeah we're going to buy bid these stocks it's like guys this happens every three months um and it's also every all these hyperscalers decide that they're going to be sold off as well when this happens is because oh my god more dilute or more capex this is so bad but i feel like i don't know i mean obviously the market and investors aren't long-term thinkers but the roic here is obvious like i don't get why they would sell off um the hyperscalers and yeah we're going to go through the same cycle again every company every hyperscale is going to raise capex by i mean gosh at least 10 to 15 percent um and i think it's it's you know you do say that microsoft has it'll be so used to be for q1 and q2 of 2027 essentially like okay yeah they're definitely going to raise like i do would agree i think 225 is probably even light i'm trying to be conservative yeah i think you have a fair conservative approach i think 250 is probably very likely probably more than that [00:12:30] Speaker 1: 50 is what uh what i am speculating i mean yeah i think if i said 250 to 300 i think people would [00:12:37] Speaker 2: say i've lost my mind so we'll keep it at 25 to 250. yeah i guess the 50 increase would be 275 or something which would be i i mean i'm not going to say this impossible but and also i think with component pricing too is a large factor of that um that's and that's what i would say that is probably the biggest bear um thesis for neoclouds is memory prices um because that is the prices are insane isn't it i think 40 of capex is memory now is that right i think [00:13:12] Speaker 1: i saw something like that there's stats going around i'm not sure what is correct and what isn't we know that zuckerberg had to take up his capex this year and is 100 due to memory pricing yeah and 16 companies locked in five-year deals with micron so memory pricing is very critical micron just locked in a 10-year deal for wafer supply so this party's not stopping anytime soon i cannot wait to listen to the taiwan semi-earnings call and see what they say their revenue just increased i think 37 year [00:13:46] Speaker 2: over year yep i think what's what's interesting too is it might even have to be in a couple years to reframe it of how we think because i think we keep thinking of things in cycles like oh it's capex cycle when i think it could just be like you said it's just a part it's not just a part of your budget that's it it shouldn't be called a cycle anymore because it's not a cycle if if if larry fink really thinks we're not building enough um and you know gerson thinks we need decades that's not a cycle that's just spending like you know a cycle in in short means it's a short-term thing whereas this is not going to be a short-term thing i don't see how this is a short-term thing i mean if you look at what [00:14:25] Speaker 1: jamie diamond said during their earnings call they're utilizing ai more they've eliminated jobs and positions and redeploy people internally because they're utilizing ai to automate stuff and they're saying they're at the very beginning of this yeah and it's like would you call paying for [00:14:49] Speaker 2: cyber security a cycle or would you call your software development department a cycle no it's just a part of the budget to grow as a company now obviously i think it becomes a cycle when there's a ceiling but no one knows what that could be so until then i mean that's not a cycle it's just a part of how you grow and in the new age of an enterprise company and here's the other thing that i [00:15:09] Speaker 1: think people are missing the big point if ever if tokens are the new oil and if every company is going to have a token budget okay what happens when everybody gets used to using tokens yep you're not scaling them back nobody is going to build out it and this is the big thing when companies build out things to make them more productive they don't go backwards so if you build out your ai infrastructure in your organization on an ai stack or you implement ai whatever it is and you're using x amount of tokens a month in four years from now or three years from now you're not just going to wake up and say all right we're good yeah we don't need this anymore you just automated your business that's never [00:16:03] Speaker 2: yep and that's exactly why i think investing in the inference layer is the way to go that's why i like nebius and also that's why i think the hyperscalers are we're gonna oh my god they're gonna be such bigger companies you know five years from now ten years from now i mean inference inference is gonna win you know are you in core weave or just never i'm not but it's i think for me what's hard is i'm already heavily allocated to data centers it's like if if i want to pick my horse i'm picking nebius whereas if i had a massive diversified portfolio i think having both is really smart but for me it's just i like to concentrate and so i'm just going to pick one horse um because i do think honestly all three are going to crush and i do think core weave at this valuation is kind of atrocious but i just prefer the long-term vision of nebius and i'd rather hold it [00:16:53] Speaker 1: than play so it's more of the vision of nebius not anything else okay yeah it's more of the vision [00:16:58] Speaker 2: how they think about it and just how they approach the sector as a whole like i think they just they navigate far more effectively i think in the sector and are more of an engineering type of team but i do think corey was a great management team and a great company um yeah i think it's just more of a division and how this guy would kind of like to invest somewhere about here i know you were [00:17:20] Speaker 1: in right i mean i've been in that forever same nebius was the first one i invested in thankfully it's done a lot better than core weave but yeah yeah i look at the numbers and and i'm not saying nebius isn't going to do good i own nebius but i look at the revenue for revenue core weave it's just [00:17:42] Speaker 2: it's it's it's getting it's getting it you have here's do you have um forward estimates by chance or just up to q1 or just pre just trailing 12 estimates i'm sorry what did you say you have forward [00:17:55] Speaker 1: estimates so i can give you the forward estimates i'm just looking at the revenue on a quarterly basis [00:18:02] Speaker 2: yeah okay cool so you want to see um yeah like what analysts are guiding for for 27 and 28. so i'll bring it up because i i honestly don't know much about i'm pretty uh i would say ignorant encore even like the numbers i know they're triple or 4x larger than nebius um yeah this is one of the [00:18:21] Speaker 1: big things that got me to really turn and be like okay it's time to really add to the position so the analyst community is saying that core weave is going to lose money this year and lose money next year but in 2028 there's 11 analysts coming out and saying three dollars and three cents of eps so it's trading it theoretically 26 times forward and then five analysts are willing to say they're going to do seven dollars of eps in 2029 puts them at 11 times wow that's crazy growth interesting that they think [00:18:59] Speaker 2: they're going to be profitable in 28. i would disagree but that's i wonder if they're already doing a [00:19:05] Speaker 1: billion of ebit on a quarter it would make sense yeah that's true they may even be profitable i don't want to really say next year but it's a possibility depending on how quickly they can stand everything up but yeah we don't think they're going to continue to build and build that's the big question [00:19:25] Speaker 2: yeah that's why i feel like with analysts i kind of disagree because i just think the bill is going to [00:19:29] Speaker 1: keep going 34 analysts are willing to say 12.6 billion revenue this year 25 billion next year and then 40 billion the following year like crazy numbers this is where i think nebbius catches up [00:19:47] Speaker 2: so obviously this year they're only gonna do three and a half next year they're gonna do nine i think [00:19:52] Speaker 1: nine to 11. bring it up for you right now so nebbius revenue 3.39 11 and 21. yeah yeah still great curious [00:20:04] Speaker 2: where's the eb are they think they're gonna be profitable too can you go up a little bit next year i think next year okay see this is where i'm like i mean maybe i wouldn't hate it that's the investor um i just don't see it but i would love i would love [00:20:23] Speaker 1: for it to happen yeah i'm just i'm at the point where the bear case doesn't scare me and everybody needs more compute we are compute constrained everything that i read is showing us that these companies cannot stand up data centers quick enough and i'm looking at the meta news as a buying opportunity because medic can just turn around and resell the compute they're getting from core weave for a profit yeah i was gonna i was nasty but i haven't [00:21:03] Speaker 2: read any of the contracts are they allowed to like is it because they do whatever they want i guess [00:21:07] Speaker 1: technically if they're paying for it why not yeah yeah maybe they can but i mean they can just sell their own either way oh yeah no i'm just curious when i first came out yeah 100 yeah they would [00:21:19] Speaker 2: definitely sell for premium i mean when i first saw that i was thinking like wait can they sell net bases in quarters compute but yeah you're right if they're paying for it they probably do whatever they want um and i think what's really going to happen too with neoclads is i genuinely believe they would go on a crazy re-rating the second there's profitability because the whole thesis is oh it's unprofitable and they're never going to make a dollar um and so like you said of core weave can be profitable next year i mean that company is going to get a crazy re-rating um far higher than 40 billion [00:21:51] Speaker 1: i'll tell you that that's for sure the only thing that worries me is the apple lawsuit with open ai you have 400 employees that now work at open ai that were at apple you go into discovery phase i get very worried about what comes out when the subpoenas start flying and you get text messages on cell phones you get emails you get affidavits i worry about what's going to come out and if that is going to impact open ai to the point where they either lose customers can't build a physical hardware device or even worse apple just 86 is them from the app store that would be yeah i i would agree that that's terrible for the capex cycle then how much new day center demand is open i really need yeah [00:22:54] Speaker 2: right very little right and that would create yeah that would that would make it a cycle at that point it would it would hold a lot of builds that's for sure that would be my bear case right now [00:23:08] Speaker 1: that i would be pounding the table on if i was a bear i would just say listen there is a likelihood that in these interviews these apple employees were encouraged to bring their laptop encouraged to bring their cell phone and then maybe they who knows but we're going to find out and discover it and i would that's the drum i would be beating if i was a bear yeah i do think that would affect the [00:23:35] Speaker 2: cycle the nice thing is nebius has no correlation to them with all their contracts and who they have um but it would still collateral damage that's yeah 100 yeah it would it would yeah that's exactly the right way to put it it would regardless of that it would they would yeah that would be bad um do you do you know the exact dollar amount that's tied up into their agreements and promises [00:24:01] Speaker 1: and all the contracts they have with everyone else i've looked into oracle it's 44 supposedly of oracle's contracted backlog for their rpo yeah oh god oh man you're talking like 300 or 298 billion dollars of backlog is tied to open a guy so if oh man i think if i remember the numbers correctly if 100 of that went away they'd still have 338 billion still on the books it's a significant number [00:24:35] Speaker 2: of times still a lot yeah yeah that's insane jeez i mean is there i guess on on the topic of bear cases is that you're probably your only actually [00:24:48] Speaker 1: good bear case you think um i can't think of another one yeah everything i've read bears i disagree with [00:24:57] Speaker 2: yeah what do you i mean what about your thoughts on just component memory pricing you just think it's it'll come down the consumer sort of mentality or what are people not gonna buy it right yeah [00:25:10] Speaker 1: so if zuckerberg decides not to buy it or that's a bad example if sundar decides not to buy it but jesse and nadella do they're gonna fall behind they don't have a choice right i think the real question is how good are the tpus gonna get that broadcom is making for everybody because everybody is signing these multi-generational chip deals and how much are they trying to get away from the nvidia tax [00:25:44] Speaker 2: yeah i do think from what i've seen it feels that feels very similar to the open source versus closed source model where they had different purposes where it's like dpus are good for very specific purposes versus gpus and same with open source versus closed source like open source isn't going to eat as you know claw's lunch but it will eat probably the very minute easy low-level tasks because it could become a lot cheaper which is what i'm reading and kind of seeing more and more is that the companies that are have these massive budget or do have these massive budgets are spending a lot of the easier tasks on open source just forking the chinese model then building their own um on top of it layer on top of it and then still using but they're not getting rid of claude anthropic they're just it's jevin's paradox they're just spending more they're just like oh wait we can use we can limit our token utilization still spend the enterprise plan on claude and codex um but we can still use open source for the cheaper level of the test oh this is great now we're even more efficient so it's not reducing the overall budget to me from what i'm seeing it's expanding it it's giving them they're realizing more and more what more capabilities they can do um and it's just it's i don't know that's sort of how i have you tpus though and the open source so i have a theory for service now [00:27:02] Speaker 1: and if they're not doing this i would love to build a company maybe we could build a company to do this whoever figures out how to be the gatekeeper as to what model you use and have all the models that you're paying for in one place and allocated per task automatically to one maximize your token leverage and to minimize cost that company is a trillion dollar idea sounds like pound deer now i don't know is that is palantir doing that maybe maybe not for me to see them not i also think that's also nebbius's approach too you know maybe palantir is gonna implement that and there'll be even more of a reason for people to get a aip in because it would be model agnostic but that's the that's the key of the future whoever is allowing that switchboard of models automatically that's the key to the future [00:28:06] Speaker 2: and that's i 100 percent agree and that's also why i like nebbius because i think they're thinking that way um and they see the red in the wall that's why they acquired agent ai because that company what they focus on is getting the most tokens out of a gpu that's why they have a partnership they just had a partnership actually with a company called um oh it's from it's removed i can't remember basically 2x google demon engineer started their own open source company and they just gave a billion dollar partnership with nebius until 2029 um as an open source platform and so i think nebbius sees that right in the wall as well and they're trying their best to attack that um but i do i would agree that i've been thinking about that as well a company that can have the 15 000 different models come in and decipher which one is the cheapest way to go i think that's 100 the route you know enterprise is going to go in the next five ten years [00:29:05] Speaker 1: yeah we're going to say i all i know is i have no clue what's going to happen i can only make assessments based off the information that we have and it just doesn't look like anything's slowing down and i cannot wait until we see nvidia's numbers now it looks like that they're exploring financial hedges to protect against future decline in memory and storage chip prices i mean who knows we'll see what happens i mean this is going to have a lot of a lot of stories but i mean [00:29:39] Speaker 2: anthropic is preparing to go public who knows by september or october right something like that yeah yeah um i dude i love that i love our in cycles it just validates the thesis every time until that one time it doesn't but i think we're a long way away from that um yeah i love listening to their the friends calls and they just come on and say all the great things to do with ai and the capix is they need to spend more and more and i'm like oh well that just validates my neocloud thesis [00:30:07] Speaker 1: i could turn it off now basically um and now apple is looking to get in the game and they're looking [00:30:17] Speaker 2: to acquire ai chip companies really yeah was that like a new headline today or recently i didn't see [00:30:26] Speaker 1: that i think today yeah and did you do you follow what apple does with their computers not enough i mean they're great computers but nothing more than that they are getting rid of the m6 upgrade line so right now they're on the m5 chips they have the m6 coming out supposedly in the fall there will not be a pro or max they're going directly to m7 because of how important the ai cycle is and how they were building the seven the headlines are the pro and the max chip for the seven will come out i believe end of 27 beginning at 28 and then the ultra comes out for the m7 and it supports 1.5 okay 1.5 terabytes of ram of ram yes oh my see i'm a pc nerd so i i can understand what you're [00:31:35] Speaker 2: saying that's 1.5 terabytes of ram yeah is it the room is that confirmation or just rumor no that i mean that's like every rumor but basically confirmed yeah yeah that's and that's crazy and the reason [00:31:52] Speaker 1: is because i work loads so you can run fable mythos 5.6 locally on your computer this is gonna be [00:32:01] Speaker 2: interesting man like if we can get to a stage where laptops right could do what we need to do then right [00:32:08] Speaker 1: how the hell does this go i mean honestly people may not think that that's a big deal he goes i haven't thought about that he goes why would you want a laptop that i can run it locally i can just run it in the cloud it's what we started talking about with sovereign ai now you can just buy a laptop at the apple store and run your workloads you don't even need a whole server room yeah see that's interesting [00:32:33] Speaker 2: i didn't even think yeah i didn't even think about that that's that would be that's going to be insane [00:32:37] Speaker 1: that's only 18 months from now that's exciting we're halfway through 2026. yeah do you think [00:32:45] Speaker 2: could that be no no i was gonna say could that be a bear case definitely not i think jevin's paradox where everyone yeah now everybody's doing it you need more hardware because now [00:32:56] Speaker 1: everybody's got to upgrade blah blah blah you talk about corporate upgrade cycles think about what happened with windows 11 back in october 2025 you're going to see the next big pc upgrade cycle it's not going to be windows 11 to windows 12. it's going to be to computers with motherboards that can handle enough ram to run ai workloads on your desktop yep no that's awesome i didn't know that yeah it's it's [00:33:22] Speaker 2: it's weird though too it's i feel like apple just hasn't been this ai race i guess this is not their thing but it'd be interesting if it becomes sort of like at home like gpu sort of i don't know it's like like they could go a weird direction with it i mean they have the ecosystem they have an amazing product i mean one that terabytes is insane who knows what's gonna be in 2028 29 too like it could be three terabytes or five or but it's like then it's then that goes back to so then aren't we early on memory so that's that's why i'm bullish dram like um we don't even have robots or full self-driving you see that's where i yeah that's where i start to like i think my brain discombobulates because i can't even think in terms of like because a robot's gonna take 20x the memory than a gpu i mean what are we talking about here um and then it's like i can't even fathom isn't micron going to do more profit this year than which company what is it all the hypers not all that not all of them combined but more than [00:34:29] Speaker 1: most of the hyperscalers did what 58 billion of operating income or something like that [00:34:36] Speaker 2: like even a year ago if you told me that i've been like shut up and now i'm like it's getting crazy so [00:34:44] Speaker 1: i think that the neoclouds are going to do very very very well i think there's a lot of companies that are going to do well and tech is going to be running the gamut and it's amazing that the legislators in new york put a mortatorium on data centers yep and you know what the best part is [00:35:04] Speaker 2: core weave and nebius and all these companies sold off and they don't even have it is that are in new york i mean maybe it's the the market is trying to see the trend and that's why they're selling off these companies because maybe this affects a bunch of other states not just new york but also like nebius and core have no data centers in new york what are we talking about here well here's the big problem [00:35:23] Speaker 1: somebody said to me on x what does it matter micron's breaking down on a fab in new york and i said well it matters because when you say no this is not the united states of new york it is the united states of america states plural keyword is plural key letter s states now you have 49 other states somebody's going to say yes somebody's going to offer them tax incentives and you know what happens when you put a moratorium on and you say we're putting a moratorium on for a year to do our environmental studies and do this responsibly they already signed a new contract they're going to break ground somewhere else and they're not going to want to deal with your bs and whatever bs regulations you decide to come up with because you are now talking to somebody who their sole purpose in life is to study soil trust find something wrong with it yeah and it's find something wrong with the same way if i drew a picture and you gave it to a real artist they would find something wrong with it they'll find something wrong but you know what you're going to lose jobs you're going to lose jobs on the initial build then you're going to lose jobs afterwards not just internally at the data centers but the supporting area around it you need food you need electricians you need plumbers need carpenters you need all the trades you need roads you need everything these people just made such an economic blunder and you open the arms for 49 other states to say pick me pick me pick me yep yeah he's a hundred [00:37:18] Speaker 2: percent all yeah all you do is open the doors for the other states be like hey we won't do that come to us and then the only the only people or the only thing that this burdens is the state it doesn't burden the companies they all i also made a funny tweet of like oh no the company that no one wants to build in isn't allowing people to build in it it's like it's like why do you think these companies were coming there in the first place because they knew something like this could happen 100 so it's just yeah it's like i i don't get it it's the dumbest thing i've ever seen and but hey they're to their [00:37:49] Speaker 1: detriment you know not to mention new york's a huge state massive you can put it somewhere where it's not [00:37:55] Speaker 2: going to impact anything yep 100 and it can be behind the meter where it doesn't affect people's [00:38:01] Speaker 1: electricity bills and all these other things but hey you know i'm glad i'm glad they're listening to people who don't have new york's best interest at heart i'm glad they're listening to people that are scared of their own shadow and don't want more jobs in the state or more tax revenue coming in to offset everybody that's leaving but hey what do i know i'm just some guy that decided to make a youtube channel [00:38:27] Speaker 2: same yeah honestly what do i know and then also of course the president comes out you know the day or two later and says what these states are doing in a ridiculous state of senator of the future they provide jobs etc etc kind of like what you just said so well i am gonna say i think president trump [00:38:43] Speaker 1: copied me just gonna say because i came out with a tweet first about it um i'm gonna find it here we go the day before the president came out with that i put this out on x and i also talked about at length on the stream so i was before him on this one my home state of new york continues to make bad policy decisions the construction of data centers equals jobs and economic output what's worse is if we get sweeping legislation in the future that nationalizes this america will likely fall behind china in the ai arms race in the 1960s we did not put a moratorium on space and let russia beat us to the moon well said i mean nothing else to say to it [00:39:36] Speaker 2: was there any uh any pushback or rallies to trying to get the moon first in the 60s [00:39:43] Speaker 1: i don't know all i know is this is so bad because if more states follow and then if you actually get look we're gonna have regardless of what party it is we are going to have in 2028 a different president doesn't matter if it's left or right if whoever is sitting in that seat and whoever is our elected officials in the senate and in the house they could go down the road of a data center ban and if they do they hand supremacy to china because you know what china is out there building nuclear power plants they're building their grid out and they're building data centers and you know what happens then they become the world's dominant factor in ai and when you fight wars in the future it's not gonna be missiles it's gonna be oh we're just gonna hack your grin take it down yep oh you have no lights you have no power no refrigeration no nothing for three days you're screwed complete chaos and pandemonium [00:40:50] Speaker 2: good luck yeah no i agree i i don't i don't get why we don't want to win um yeah i honestly i can't even conceptualize i don't really i really just don't understand what what the point of it is at the end [00:41:07] Speaker 1: of the day like i just i don't get it i just want common sense it's common sense that's all i asked for i think that's what it is i just want common sense i could care less for too much steve i want too much like one plus one equals two yep that's it one plus one equals two yeah no i agree i mean you gotta laugh you gotta laugh so heading into earning seasons let's end on this what are you most excited for this earning season we just got the banks so you know text next yep man i think for [00:41:50] Speaker 2: me um i'm actually really curious to hear what i mean i think it's honestly the same every cycle now obviously nebius and the the neoclouds just to see what they're all saying and kind of commentary i'm honestly really curious i think the most one i'm excited for is probably meta because i feel like zuckerberg's really trying to go down this open source approach and so i'm curious to hear his thoughts is like we've seen more and more talk of the town of open source and kind of that route and he's trying to combat it seems like anthropic and open ai um yeah i'm just i'm curious to see because i think a lot of the answers questions are going to lean towards open source and open source and open source i feel like that's kind of a lot of the questions are going to be asked so i'm curious to see how jassy and you know satya nadella and senator bachai and you know zuckerberg all answer and kind of think about these questions um and the way they're trying to approach things because it's it's sort of a they have to be careful um with how they answer because they have massive agreements with the likes of open ai and philanthropic and it's kind of muddy waters there so i think it's always the same though it's always the four hyperscale or the well we can maybe call meta hyperscaler now the four hyperscalers um with nebius and core weave and obviously iron just curious to hear what they have to say and then honestly just in terms of ai in general love hearing what palantir has to say um i love hearing also just tesla commentary about the human robot curious to see how that's progressing what's going on there um but yeah anything ai and just the progression and what people think in commentary there is what [00:43:25] Speaker 1: i'm excited about the only thing i'm interested in from tesla is optimus yeah same everything [00:43:34] Speaker 2: yeah they could get rid of the car business i wouldn't even blame honestly um obviously i get why they need it and it's trains the bot they have fsd but i i am way more curious on optimus i think that's that if it really does work out that will 10 to 20x the company again [00:43:53] Speaker 1: well i'll tell you this i am most excited for microsoft right now as much as i want to see what zuck says i care more about the cat box number out of microsoft because that sets the tone for meta alphabet and amazon because they're not going to release their new numbers until [00:44:11] Speaker 2: next year yep and then i'm excited for nvidia to come in at the end of usually august to kind of come in and shut up the bears with another like 90 growth quarter at a three four trillion dollar market cap um and continue to be a trade at a disgusting valuation um for how fast they're growing and how much profit they're raking in but i'm also curious to hear about they just released and they just released like a new like um revenue sharing model as well here's to see what jensen talks about how he thinks about that [00:44:46] Speaker 1: um but yeah it's gonna be interesting to see what all these ceos say and putting everything together at this point there's no nothing to make us think that this is slowing down anytime soon and if core weave keeps falling if nebbius keeps falling i am a buyer until there's a reason for me not to be bullish and i'll change my mind if i if the data shows it but when you're scaling revenue the way core weave is and you're now doing a billion a quarter in ebitda and the interest payments are what they are i'm not really that concerned especially with the amount that these data centers are or these companies are spending on data centers and the fact that alphabet said we would have shown more revenue if we have more compute whatever they stand up they sell [00:45:45] Speaker 2: yeah i mean even nebius said last quarter for every gpu we have we have four to five enterprise customers fighting for it yeah i mean it's enough said like hello people are asking meta for their [00:45:59] Speaker 1: compute i mean it's just going to show you zuck is going in that direction for a reason yep 100 we'll see all right any other topics you want to get on or oh not really okay um yeah that's [00:46:14] Speaker 2: all that's all i have i think the only thing i want to say is uh i the only thing that i can't stand is the iron management team that's all i gotta say wait the what the iron management team oh wow [00:46:26] Speaker 1: listen sure let's just give them five percent for no reason it was really great with no incentives or anything i just give them yeah whatever i didn't have a problem it was the type of package elon got [00:46:40] Speaker 2: where it's all performance based yeah same with sofi anything you know just anything performance but even if it's a crappy metric just anything anything to just hand them the money like what are we just [00:46:51] Speaker 1: ridiculous they can legitimately just sit there and do nothing for five to seven years now yep they don't have to hit a single metric it doesn't matter they can eat donuts if they want and that's where [00:47:02] Speaker 2: it's just like that's why i'd rather pick nebius or core even though i think everyone could do well just because they can sit in their hands and do well because competes are scared sure but it's wrong [00:47:11] Speaker 1: just to have a carte blanche blanket deal like at least look noto he didn't get two out of the three tranches he only got one of them awarded we have no idea if elon's going to get his tranches awarded yeah didn't he just get the 25 dollar one yeah yeah that's the one he got the 25 he didn't get the 35 or the 45. yeah yeah so yeah it should be performance-based because you know what they should do well when the shareholders do well 100 what do you think it cares what do you think of the the logo on the jersey the 50 million a year for that do you care i think it's out of out of control there's no reason i agree why are they why do they need to spend that money on that [00:47:57] Speaker 2: why yeah i i agree too and i think people tried to correlate it to oracle or salesforce like guys here's these are established a hundreds of billions of companies that have a software layer [00:48:08] Speaker 1: when when sofi did the sofi stadium deal noto specifically said i have a marketing budget i can either spend on ads on tv or i can buy naming rights at a stadium that shows up at every concert that's played there in every football game that's played there and it is sofi stadium and that gets ingrained in everybody's head yep or i can spend money on ads he chose to do the naming rights iren does not have a consumer-facing product it doesn't benefit them it benefited sofi because they have to spend on advertising either way to grow the brands that was a business decision on where is the money best spent they're going to spend the money anyway iran has a handful of customers they don't they're not selling a product to you and me nobody needs to know their name that was a vanity thing and another reason why i'm like [00:49:10] Speaker 2: this company's dumb yeah i don't get what they're doing that's all i had to say though those are the two things i always wanted to dish it out it doesn't make sense no not at all cool glad i don't feel [00:49:23] Speaker 1: insane feeling that way but no i think it's ridiculous all right nate i appreciate you coming on we'll get together once uh everybody reports and we'll see what happens and if our views change so thank you very much please go follow nate on his youtube channel and his x account i'll put the links in the description below please go follow him please go follow him please go follow him please go follow him please go follow him please go follow him please go follow him please go follow him please go follow him please go follow him please go follow him and we'll see everyone later have a great day have a great day [00:49:51] Speaker 2: Love Steve. [00:49:53] Speaker 1: Peace everyone.