원문 보기
Jam how do you how do you how do you look at this I think that value and money are correlated in some Industries like real estate or heavy industry but I don't think that it's very correlated in Tech and I tried to say this in my tweet I don't know Nick can you find this tweet where I was just talking about the Deep seek model so what did we see last week we saw a model that got released by the Chinese under by the way a totally you know open source MIT license that you can run on a laptop and it turned out to be as good as open ai's 01 model so a couple of versions old but the point is that this model cost millions of dollars and was able to compete with a model that took billions of dollars to train so my take away from that was that these costs are falling really precipitously so that's sort of my one observation is spending more money doesn't necessarily get you further along if the goal is progress and I think that that's very true in what we're learning about AI because we're learning as we go along so then I think it just brings up this question of what is the point of saying all these grandiose numbers and this is where if you look at a different industry like if you if an oil company came into the oval and said they were about to spend 500 billion you kind of know that it's real because you know what these things cost and you know what the entire infrastructure will be and you know it's so we know we know what goes on yeah so the deeper the oil is in the ground and the harder the rock is that you got to break through to get there the more you have to spend so those numbers become very grounded in reality if you were a rocket company and you wanted to build a rocket and send it to Mars you can also be very detailed in what the costs are and you can assume some amount of reductions of scale if somebody said you'd have to spend 500 billion the thing is that when in this world it seems that there are these an es that are happening that allow you to do things cheaper and cheaper almost overnight by spending 1 1000th of what you spent even 6 months ago I think the spending money part to be honest is more of a gimmick than it is a technical commitment so I guess if they want to spend the money go ahead but it doesn't actually tie to whether they're going to be able to make custom vaccines whether they're going to cure cancer so I kind of thought that these things were just join and so my takeaway was entirely different it was not a commentary on Mosa or Larry or Sam I think all of those three companies are frankly very good it was more a comment that you have to be very careful to protect the president's Legacy if I were them to make sure that the things that get announced are actually further down the technical spectrum and are actually going to be real because if they achieve these things but it costs you a billion dollars and you only hire 50 people there's going to be a little bit of egg on the face and so that was sort of my own takeaway I think that the things were decoupled it just seemed more marketing and Sizzle and kind of hastily put together it was definitely yeah put together you mean yeah you got that sense but I think it would be great if if open AI builds another incredible model whatever comes after 03 04 05 but it's not clear that you have to spend 500 billion to do it you know I think people are playing to what Trump likes which is a big number and which is his influence on growing the number so there was specific details where Masa said you know you I I told you 100 billion you challenged me to get two I came back with five right and so this is the sort of trump ethos is hey we're going to swing for the fences in a way you know you could be critical of it and be like well Trump's a BS artist or whatever or if you're from Silicon Valley you could say Hey listen he's moonshot it he's going to just shoot for the stars and get the moon so if these guys wind up spending 50 billion or 100 ion it makes us more competitive more power to the question I had was you know was did Elon have any insight that this was happening probably not from from the reaction on I bet he did you know I think you're I have no Insight either but I do think if you were going to be critical of trump that he was putting his thumb on the scale for any one company or individual this shows that he's not it's an open platform the the Trump Administration it's an open platform highest bidder gets to come to the White House whoever spends the most whoever buys the most military you know buys the most ammunitions from us you're going to get your time in the white house it's uh this is a capitalist country and the greatest one in the world equal opportunity recently inaugurated president Donald Trump has teamed up with industry Giants Oracle open Ai and SoftBank to announce a massive 500 billion AI initiative known as the Stargate project as anticipation builds questions Loom about which companies will emerge Victorious from this colossal investment among the leading names Nvidia stands out as a key player likely to benefit profoundly from this project but could this be the investment opportunity of a lifetime chamath popatia think so welcome back everyone to invest ocracy okay so there's three points I want to make here go Point number one going back to the origins of the of this podcast and you know sitting watching both Public Market investors and and Venture investors you know one notable difference is you know I think Venture investors tend to be very tribal so if you're in One Tribe You Are by definition not in another and you know I think there's a lot of kind of domain specific reasons for that you know for me as my legacy as a public market investor I'm very comfortable being both you know Pro open Ai and pro Tesla and Pro X and so I just kind of cave at that kind of you know as a first point my second point is can we do a slight detour on masa for one minute and I'm curious to get you guys's opinion but is Masa the goat because when I started in the growth business I met with Yuri mner and he said Thomas you get in the Hall of Fame if you can have one deal that gives you a billion dollar return I.E you know whatever you invest plus what you take out is a billion dollars or more and he actually kept a Tracker in his mind of who had made those deals and what share he had and Etc so let's go to masa for one minute Masa is the hundred billion doll Club yeah okay and not only that he's actually done it twice he did it the first time with SoftBank and Alibaba which at its peak was a $200 billion win and I think he ultimately netted like 80 plus billion out of it so that's one he did it again with arm that he bought for 30 and now is sitting on a gain of 100 I think 35 billion and he could have done it a third time with Nvidia if he hadn't sold I mean guys there's nobody like him there's incred incredible you're 100% right you got to give the man his flowers he is the so I'll tell you what chath I am never B against Masa ever because you got to give the man credit for what he's done and so now let's go to kind of open AI you guys mentioned we we do own some open AI I am bullish on open AI which I guess is now a uh counter Against the Grain kind of view but guys my point of view on open ey and the reason I'm so bullish is is really kind of chat GPT and I do think when open a gets you know we do conf it does have different businesses right it's got the model business it has the API business and then it has the chat GPT business right and if you look at the downloads maybe Nick bring up the charts but it is really interesting to watch the market share of Chad GPT versus Google uh Gemini and grock this is both in the US and international it really is a dominant kind of franchise right it is maintaining % plus market share you know J Cal if this was right sharing right we would say wow this this company is the winner right so to me I think the the quality of the chat GPT franchise right I think they disclosed 300 million weekly actives over a million Enterprise users it has become a core part of my own workflow you know I have it on my phone I have it on my desktop I use it all the time I can't even quantify value that it brings it's so important for what it does for me so and I do think that it just keeps getting smarter the Stargate project dubbed the largest AI infrastructure project in history by President Trump has captivated the tech world and investors with its promise to revolutionize the AI landscape Oracle open Ai and soft Bank have already committed an initial $100 billion to this new Venture with potential investment reaching $500 billion over the next four years the Project's vision is clear to establish American leadership in AI by constructing cuttingedge infrastructure across data centers accelerated Computing systems and more envidia is poised to be a significant beneficiary of this AI Revolution with gpus like Blackwell in high demand and Tech behemoths such as Microsoft and Oracle already on board nvidia's inclusion in Stargate is hardly surprising on Stargate specifically look I think the real question to me isn't whether they have the 500 billion right because we know that the 500 billion could come from Equity it could come from Masa if you look at Gavin's tweet by the way he does caveat it right saying that assumes that Masa doesn't want to sell arm and soft bank right so I'm not sure just Le up and use debt correct so you have debt I I think to me the real question guys is do we ultimately believe that you can get an Roi on that 500 billion because if you can there will be people that will want to fund it right you can do a data center by data center you can do it actually at the GPU level right so the financing is there to me the the 500 billion doesn't scare me from an absolute number I think the question we you know really have to ask ourselves is do we believe that the market is big enough that the opportunity is big enough that investors can get a return on the 500 billion I think I think you so to me that's the real question I think you framed it perfectly and since you bring up ride sharing and the 80% there I you know I really think that's an important thing to to take a moment to um to settle on because in that case you didn't have Uber being challenged by Google Microsoft and Elon Musk you know at the same time in a digital product as opposed to a real world product and so you know when I use gemini or xai and chat GPT and I use them all daily and Claude the differ the difference between the products right now is really narrow and I think think you know if the equivalent would be somebody who had a 100 million cars on the road already competing against Uber which you know doesn't exist or didn't exist at the time so that that I think is the the way I look at this as a slightly different type of race I do believe the chat GPT franchise has some is worth something but I think Enterprise folks when I talk to them want the open source product most of all so anyway I think it's still anybody's um I think it's anybody's game right now and I don't know how you get a return on 500 billion invested in this that that's the other thing and I don't know if that's the real number free BR what do you think I don't know yeah this is out here in some crazy guys the way these things happen right is you don't you don't fund it all Equity UPF front right I think these things get funded Facility by facility data center by data center across four years right and to Chamas point it's not just Equity it's you know sight level debt and Etc right it's a complex capital structure to these things the necessity for top tier gpus to drive the Project's buildout presents Nvidia with a potential Gold Mine further boosting their already impressive Revenue trajectory picture this with a single Blackwell GPU priced between $30,000 to $40,000 and systems requiring multiple gpus the financial implications for NVIDIA are staggering in the explosive $200 billion AI Market projected to surpass a trillion dollars by the decades end the Stargate project sets the stage for NVIDIA to shine even bright obvious consequence of this data center and Chip infrastructure effort which is obviously going to happen with or without government involvement is an increase in electricity demand and Thomas I know you've got some slides do you want to pull those up and and show them because I think they're something I've talked a lot about on the show which is the difference in uh electricity production capacity in the US versus China historically and then prospectively going forward ultimately as I've talked about in the past right the us today is paying roughly you know call it 1 and a half to 3x the price per kilowatt hour for electricity in the United States over what China's paying we have I believe half of the electricity production capacity of China they're adding more the cost to add is is on10th of what our cost to add is call it on10th to 1 half our cost to add per new gigawatt of production capacity everything is in the wrong direction and ultimately if Ai and automation are the critical factors for economic growth going forward and electricity is the critical input to that we are hugely disadvantaged and aren't going to catch up seems like you've pulled this analysis together which kind of indicates the same if you think it makes sense to walk through this now yeah yeah I'd be happy to I mean Dave you're 100% right if you think of the GPU as the atomic unit of the future right and and the key competitive advantage and and the key kind of uh ingredient you can't power a GPU without a Data Center and you can't use a data center without power so I do think at the end of the day it really comes down to the grid and it comes down to power in conclusion the Stargate project Heralds unprecedented growth opportunities for NVIDIA fortifying its dominant position in AI infrastructure with an expected Revenue Boon from this ambitious project nvidia's stock continues to present as a promising investment but will you seize the opportunity to go all in and invest in Nvidia
Comments 0