SpaceX Files for Nasdaq IPO | Bloomberg Tech

Added: 2026-05-25

[00:00:02]Bloomberg Audio Studios, podcasts, radio, news.

[00:00:12]Bloomberg Tech is live from coast to coast with Caroline Hyde in New York and Ed Lllo in San Francisco.

[00:00:23]>> This is Bloomberg Tech. Coming up, SpaceX pitches a $28.5 trillion opportunity from AI to Mars ahead of an all-time blockbuster IPO. We break down the S1. Plus, Nvidia fails to reignite the AI trade as CEO Jensen Wong pushes to diversify the chip maker, saying AI is set to go mainstream. And another major IPO on the horizon. Open AAI preparing for a filing which could come as soon as tomorrow. Let's get straight to our top story. SpaceX has publicly filed for an IPO and in the process disclosed billions of dollars in losses along with a super voting share structure that would give Elon Musk sweeping control over the company. The ticker will be SPCX.

[00:01:11]The filing also underscores just how ambitious SpaceX believes its future could be. Today's big number, $28.5 trillion. the total addressable market the company says it's pursuing according to the documents. Let's get more on SpaceX's IPO and on Starship at the heart of it. Bloomberg managing editor for Space Benedict Camel, Bloomberg deals reporter Ryan G. Benny, I start with you. We learned a lot about the reality of where SpaceX's business is right now. So, let's start there. The numbers that jumped out of you as we poured through the document and went through it with a fine tooth comb. Yeah, I mean you've already mentioned some of them and the one that really was was a bonkers number. Let's say it is the total addressable market. Um the market they're targeting just shy of 90 uh 29 trillion. So that's one number that I think everyone sort of gulped at. Um but also on the other uh sort of end of the spectrum, the the sales and uh and and the loss are actually not that great. Uh so you you've got a company here with uh quarterly sales of just shy of 5 billion and with a loss that's not that dissimilar. So for a company with those kind of numbers to seek uh a valuation of 2 trillion that is obviously pretty steep um we we then saw what Musk is trying to sort of retain the control he's trying to control uh he he wants to retain in the company about 41% of control that he wants to keep. So that obviously gives them a a really hard uh sort of control of the company.

[00:02:44]So he's the man in charge and he will remain so going forward. The debt pile $29 billion. I want to get into how this IPO is going to work. Ryan, what do we confirm about the banks? Who's lead left? Who's lead and who's behind? But also like the roles because if this is going to be the biggest IPO of all time, dollars raised or valuation doesn't matter in your world. does matter who's trying to pull it off.

[00:03:10]>> It does, Ed. And you know, there's been so much said and written over the past what, 2, 3, four, 5 months around who's actually going to be quote unquote leading this deal. I mean, if you look at the S1, Goldman Sachs is in that plum left position. To the right of them and Morgan, Stanley, Bank of America, City Group, and JP Morgan, the the last three sort of in alphabetical order as it were. But, you know, I think if you're Goldman Sachs, you're looking at this and saying this is a great marketing win for us. you know, we are the pre-minent franchise for IPOs in technology on the street. And then you're Morgan Stanley who maybe is saying, well, you know, we've done a lot of work for Elon Musk over the years. Um, what does that mean for us? Well, if you actually dig down into the filing, you can see that Morgan Stanley does have the stabilization agent role, which I will say, you know, is not something to, you know, really pass over. The stabilization role is crucial to trading that very first day flow. And in some cases, that could be up to 20 to 30% of that flow on on day one. um and they're obviously going to take a certain cut of that revenue. So um you know it's it comes down at the end of the day to sort of marketing versus economics. I mean our understanding is that the economic structure in terms of who's getting the fees is roughly equal but if you're Goldman Sachs you're able to go out to your investors and say or your shareholders and say um you know we are leading the biggest IPO of all time and that's clearly huge.

[00:04:26]>> And we we got the reporting that David Solomon Goldman CEO slid into M DMs to get that win. Uh Benny, the other important thing is who controls this company? What do we learn about Elon Musk's voting power?

[00:04:37]>> Yeah, so he retains absolute control over the company. Um overall about just north of 40%. Um some other interesting investors in there. We're not quite sure what the role of Google is. They're obviously an early investor in this, but this is a company that will mint uh lots of billionaires and possibly the first trillionaire if Musk gets his way if the pricing turns out as hoped. and he could really get into a very different stratosphere altogether if he gets those bonus shares. Uh that's something that was also teased yesterday. If you manage to actually put people on Mars, if you manage to inhabitate Mars uh with a colony, I know this all sounds, you know, deeply science fiction, but if he does that, um and if he takes the valuation of the company to a certain level, he stands to gain another 1 billion in bonus shares. again uh that will will sort of lift an already very wealthy man into a very different stratosphere.

[00:05:29]>> Bloomberg's Benedict camel and Ryan Gold, thank you very much. The SpaceX perspectus asks investors to believe in three things. AI in space, a million people on Mars and Elon Musk's ability to make it all happen. But that's also what great IPO storytelling is supposed to do. Sell the investors on a future that doesn't yet fully exist. Joining us now is Lauren Webster, Piper Sandler's head of investment banking. Just state off the bat, you know, Piper Sandler is not involved in in this IPO, but you know this industry, you know how an IPO works, and this will be the biggest IPO of all time. When you read that document, as far-fetched and out of this world as it was, what were you thinking?

[00:06:13]H I I was certainly thinking aspirational, visionary, but let's unpack for starters the TAM that that you spoke about >> that that okay so you have this document and you have to say something and what you say is we see a world where the market ahead of us is 28.5 trillion but 26.5 trillion is just AI.

[00:06:35]>> Yep. Yeah. So so two observations. First um I think every perspectus has a very lofty tam in it. So this is not out of the ordinary. It truly is par for the course. Um the second one is uh every company going public wants to be able to tell investors we have multiple ways to win and multiple TAMs to be able to touch. So yes, huge numbers. We can kind of quarterback uh Monday morning quarterback what's right, what's wrong in there. But it is saying to investors, we have multiple paths here to really meet that growth outlook that you're investing us to achieve. Is it believable the TAM number any of it?

[00:07:16]>> I I I don't know if I've fully truly believed any TAM put in a prospectus and I've always been able to poke holes in it. I think probably the one that was the biggest gap for me is really that enterprise application wedge in it. So you saw a very sort of rational stairst step. The enterprise application one needs a lot of unpacking, but I think that's part of the story here is that there are really three core pieces and each of these are going to be foundational to the future of how enterprises deliver services to their customers.

[00:07:45]>> I think what SpaceX is arguing through XAI which is part of the story it owns XAI is that they are going to take on Europe open AIS anthropics Google's in the market for aentic AI. You know, they talk about how white collar work will be transformed, but is the TAM even from a potential investor's perspective calculable? Whether you believe it or not, you know, there has to be some math involved on on how SpaceX would get into that market.

[00:08:15]>> Yeah, look, I think I think they're already delivering in some of those areas. Um, certainly on let's just take the XAI piece and what they're doing with the Colossus data centers and now having Anthropic come in as a key customer there. Um but but that is a market that we are going to hear over and over again throughout this year. You mentioned OpenAI IPO coming up potentially um uh really large TAM numbers uh because it really does touch every part of the enterprise ecosystem and then flowing down into consumer daily life.

[00:08:49]>> We maybe should have started here but why does the preliminary prospectus matter? What is it that that needs to happen? you know in convincing the market to participate in this IPO.

[00:09:02]>> Yeah. So good question and a few areas where it's really uh really critical.

[00:09:07]One is the storytelling and I think unpacking that story there's three core parts that we've talked about in terms of what are you investing in. you're investing a core space business really the foundation of SpaceX launch rocket where they have a tremendous leg up on the competition um and have really bent the cost curve associated with the future space economy two is really uh telling the story around um autonomy and connectivity for that autonomy at the edge with Starlink and that we as Starlink are going to be able to empower power the connectivity required for everything we're going to do with autonomous vehicles, robots, go down the list. And then three is really the AI piece. So that's the storytelling. And then of course you also have the numbers behind that risks associated with it and informing investors of what they're getting into. But first and foremost telling that story.

[00:10:01]>> Prior to the offering, Elon Musk has 85.1% voting power based on the structure of the shares. historically in big IPOs, how has that been a factor?

[00:10:14]>> Yeah, so you've certainly seen across the tech end industry these, you know, dual share classes where founders retain uh significant if not nearly total control in terms of >> super voting share class.

[00:10:25]>> Exactly. Um I think for Musk in particular, he has a very long leash with investors in terms of capital intensive projects. uh in part because what he he's delivered with Tesla, what he's been able to do in the space economy today. And so while you know there is some potential caution, it's not unexpected that he would retain this type of control and we've certainly seen it uh in prior tech IPOs.

[00:10:52]>> We just have about a minute left.

[00:10:55]>> What do you think is going to happen?

[00:10:57]How do you think this is going to go?

[00:10:58]Draw on your experience.

[00:10:59]>> Yeah. Yeah. So a few things. I I'm really looking forward to the Starship launch this evening because I think that I think it is important. It is not binary. What it does is you're we're already seeing tremendous investor enthusiasm around the SpaceX IPO. Uh I think that enthusiasm holds it. Even if Starship doesn't go up, Musk and and SpaceX have primed us uh to understand there is failure when you're doing big things. But if it goes well this evening, I think that creates even greater enthusiasm and sort of creates the crescendo around this opportunity, which is of course what he's hoping for.

[00:11:38]>> Lauren Webster of Piper Sandler post SpaceXs S1. Terrific. Now coming up, by the way, the world's rich biggest company, most valuable company reported earnings last night. You may not be aware, but we got Troll Prices, Tony Wong, to break down Nvidia's earnings results and get a real investor's reaction. That's next. This is Bloomberg Tech.

[00:12:12]Reality check for Nvidia. Shares now down 2%. We've been flat for most of the pre-market. We kind of had choppy trading swinging between modest gains and modest losses early in this Thursday's session after actually a strong forecast and a pitch that Nvidia is diversifying from the data center giants. Jensen Wong said on a conference call of analysts that from robots to rover taxis and every kind of startup in between. Quote, we got it all covered but investors have become harder to impress. joining us to break it all down. Tony Wong, portfolio manager of the TRO price science and technology fund and across the firm um in different guises, you know, TRO a big investor of Nvidia's. Okay, let's start $91 billion plus or minus 2% in the the second quarter. It was like almost the whisper number for the the sell side and the buy side. Um but but not enough, you know, and why is that? Hey, look, I think that uh you the growth at this scale is just unprecedented. I don't think the market's ever seen anything like it. And I think you think about traditional semi-investors. Um you know, the playbook is to sell this type of growth, right? Because it's not sustainable. You know, the margins are have peaked. But um I think if you really look at what's going on at the driver of the driver like what's going on in the end demand I think it's like really phenomenal in terms of agentic AI really taking off the time of work or like time of task is expanding instead of being one shot or the you know the the agent working for a few minutes I think we're going to go towards like months of work and that's requires like a lot of compute um to have an agent go off and like complete a task ask and think and be persistent.

[00:13:56]Um, and then this other thing is I think that scaling laws continue to hold like the frontier models are getting better.

[00:14:01]The more compute you throw at it, the better they get and you actually save money by being on the frontier because it doesn't go through so many rabbit holes uh in terms of trying to complete your task. So I think if you look at like the end demand, it's actually very encouraging and I think we're climbing a wall of worry here as a result.

[00:14:18]So, so, so the thing that I was trying to digest throughout the call is with the data points that Jensen Mong and Colet Crest put forward as evidence of that, right? The scaling laws and the dollar per token and they go to this $1 trillion figure which I think they clarified is Blackwell Reuben and it's for calendar year 2025 to 2027 and it's basically a backlog, right? But the the main overarching point is that they would argue that Nvidia's own growth is trending ahead of even the the hypers scale capex growth. How do you model for that? Like is that the math that you would do?

[00:14:53]>> Yeah, one of the things that was really encouraging I think of the quarter is that I actually broke out like hypers scale versus the enterprise and sovereign. And so you know I think the bare case is that you know Nvidia can't outgrow the hyper cap hypers scale capex numbers but there's all this new TAM that's emerging um that's not hypers scale and you know I think they continue to do well in hypers scale but there's also new areas of agentic enterprise adoption I think you look at talk to financial services firms they're adopting uh a lot of a lot of this end demand that Nvidia is producing and so I think that plus robotics that's on the on the um you know to come in the future I think that you know allows s them to see, you know, not just not be bracketed in this like hypers scale kind of compute um, you know, paradigm here. Uh, Tony, slightly unusual situation, but I spoke to Jensen at some length on Monday prior to earnings, but actually a lot of what he said about the the demand supply equation was very uh precient, salient.

[00:15:52]Just listen to what he said.

[00:15:53]>> We have the largest supply chain in the world. Our partners have done a great job securing supply for us and so all of the pieces go together. The silicon photonics is lined up. Everything is all lined up. It's just that the demand is much greater than the overall capacity of the world.

[00:16:12]If you are a student of technology history, finding yourself in a situation where demand is vastly outpacing your ability to supply even in a world where supply chains are doubling or quadrupling perom, which is what he went on to say, that's been an enviable position, but right now it's not moving the needle for this stock.

[00:16:32]>> Yeah, I mean I think that you know the market is looking at relative earnings growth and where the bottlenecks are. So capital's flowing there. Um but at some point you know you look at the multiple for Nvidia it's like really attractive and in addition like there's there's a point where um you know for the growth it is like tremendous looking at the PEG ratio the expectation so if the growth is durable um you know I think you can see multiple expansion actually from these levels and that's pretty exciting for for holders and in addition like it is the platform that a lot of this AI um inferencing is going to be built on. And I think inferencing is going to be much bigger um than training and training is also still growing because of scaling loss. Um in addition I would say that like at this valuation you know they are doing a capital return program and it so I covered Apple and Nvidia uh for Tro and I was super lucky to get that I saw like you know what Apple did with the return program you know it didn't like you know rerate the stock from day one but over time consistent capital return at that valuation really expanded the multiple I think with Nvidia it could be a similar story where um it becomes less cyclical more durable as a result um so I I think you know, they're doing the right things here with the disclosure as well as like, you know, with the capital return program and I think it's just a matter of time here.

[00:17:52]>> I think right now Nvidia's trading at like 22 times forward 12 month earnings, right? I think I I'll double check it.

[00:17:59]Something you just raised is really interesting which is basically use of capital. So the frustration with Apple is like do something and maybe that that story is changing with Apple in in the the hand over to John Turners, but Nvidia has gone out there and invested in the the the ecosystem in small increments. 2 billion here, $2 billion there. How how does Tony Wong read that strategy of of investing?

[00:18:25]>> Yeah, I think it's really smart. I mean I think that when you're at the frontier of technology you have to build the ecosystem and bring up the supply chain bring up the partners and you know Nvidia has been really smart in terms of using their free cash flow to extend that and also builds their ecosystem as well as furthers the you know the technology frontier. So I think it's it makes a lot of sense and you look at the deals they've done in the private markets they've been pretty good. So like great for shareholders great for building the ecosystem and they have excess cash still. So um I think that just goes back to like they are like this single architecture platform and there's tremendous leverage on it. You see where the margins have gone and so that's been their whole whole you know bet from like day one I think is that they're an ecosystem uh company and I think you're seeing it play through here as they build it out continuously.

[00:19:14]>> We actually have a VC on later in the program whose portfolio companies you know are so tied to Nvidia and we'll have that conversation just to close the loop. By the way, Nvidia does trade 22 times forward earnings, but historically it's like nearer to 34. Jensen has this equation, more compute equals more uh tokens equals more revenues. Do you see those revenues coming out the other side for all of Nvidia's customers?

[00:19:37]>> Absolutely. Yeah, definitely. I mean, you look at like the cloud demand, you look at uh pricing for GPUs. I mean, these are like old GPUs. Pricing is holding up phenomenally well. And uh you know you look at like the enterprise demand inflection uh that's happening and you see it you know we see it in our own company we're developing AI solutions and it's eating up a bunch of tokens where you know we see the promise obviously there's experimentation uh I think you're seeing that like come through >> just real quick like on on on the socials that's what people just go nuts over right H100 H100 pricing right now like are you tracking that? Oh yeah.

[00:20:16]Yeah. I mean I think you think about the bare case um you know 24 months ago it was that like oh like the obsolescence of Nvidia GPUs is like 3 four years.

[00:20:25]Well well first of all the warranty on these things are three or four years and then also you know you have that depreciation and so you're able to like adjust uh the pricing you know but even so like the demand's been so good that um you know you're seeing more value coming out of the older GPUs and even A100s people are using them. Even A100s, if you're out there running workloads on A100s, give me a call. I want to know.

[00:20:48]Tony Wong Tro Price back on BTE. Thank you so much. Now, coming up, we're going to hear from Airbnb CEO Brian Chesy, actually on the company's use of AI models from China. That's next. This is Bloomberg Tech.

[00:21:12]The CEO of Airbnb has defended his company's use of opensource AI models from China. Last month, a congressional committee asked Airbnb for more information about its use of these models as part of a broader investigation. And the committee says it has concerns about the implication for Airbnb's American customers. Here's what Brian Chesy told me in response to that yesterday. We are not a customer of those companies. We're using a variety of different open source models. I just want to say a few things about this matter. Number one, data and privacy is most important to Airbnb. I think anyone who knows Airbnb knows we have a outstanding track record on data security and privacy. Number two, um all of our data is vaulted. No company has access to our data. It is vaulted in Airbnb. Number three, the congressional committee has reached out to us. They said um they have some questions. We said we want to cooperate and so we're in direct contact in cooperation with them.

[00:22:08]>> Airbnb CEO Brian Chesy in a conversation yesterday. How's everyone doing out there? Let's take a quick check on the markets and it's taken a little bit of time to kind of find the story. NASDAQ 100 is now down half a percentage point.

[00:22:20]The socks is basically flat. Nvidia is now firmly lower, down 2%. But actually, funny enough, Walmart in its earnings is the biggest points drag, at least on the NASDAQ 100. We have a lot more to come in the program about Nvidia, about the historic SpaceX S1, and a whole lot more. Stay with us cuz it's halftime in the program. And this is Bloomberg Tech.

[00:23:01]The world is rebuilding computing for agentic AI and robotic physical AI.

[00:23:09]Nvidia sits at the center of these transitions.

[00:23:13]>> Welcome back to Bloomberg Tech. That was Nvidia CEO Jensen Hong speaking on the earnings call from fiscal second quarter financial year 27 earnings. One of the big takeaways from Nvidia's earnings was the rise in demand beyond the hyperscalers. Bloomberg intelligence writing that it was a more important signal for the future than Nvidia's signature beat and raise the author of that research Bloomberg intelligence is Kung Jen Sabani here with us.

[00:23:40]What is the world outside of the hyperscaler? What are we talking about?

[00:23:44]You know, some people would say, oh, this is about companies big and small building on prem, right? And that Nvidia is going to have this whole new customer list that is that of that size. It's hard to see. Give me the full thesis.

[00:23:57]Yeah. You know, that second segment is very fragmented, but people keep missing because it's very diverse. So, first of all, it has So, right, which is itself driving billions of dollars of revenue.

[00:24:08]All of the AI neo clouds. So when we look at the space in the AIO clouds that keeps on growing higher and higher and they are becoming now a significant spender of capex which does not get captured in most of the capex discussions from the public clouds that we hear you have all of your AI labs uh up and coming enterprise and industries deployment of on-prem AI as well as these data center u you know guys who run the data centers are now becoming much more high GPU based AI data center cloud renters. So this 50% which is expected to grow faster than the top hyperscaler segment by the way is getting more and more significant especially for Nvidia because this is where you have the least competition. Uh the A6 and the GPU the TPUs and the traniums don't compete in this segment.

[00:24:56]So it's purely monopolistic for Nvidia.

[00:24:58]Nvidia gets the highest stack attach rate means they can sell a lot more of their networking and their other components here and they get the best margins in this segment.

[00:25:09]I think during the call there was an attempt to clarify that that kind of mysterious $1 trillion figure. So so it runs calendar 25 to the end of 27. It is Blackwell and Reuben systems but it excludes various CPU and some of the other networking components you were just talking about. I always come to you always have your feet on the ground, right? I it doesn't really matter what the number is that hits on the the second quarter 91 billion plus or - 2%.

[00:25:38]That number the 1 trillion give me your kind of like base level understanding of it.

[00:25:44]>> Yeah. All it includes is just the Blackwell and the Vera Rubin the full stack that is going to start ramping in 3Q. Uh that's all it includes. It does not include like you mentioned the 20 billion CPU standalone that they mentioned right for 26 and 27. It does not include the grock based LPX server.

[00:26:04]It does not include the new launch of their storagebased CPX servers and a few other networking components. So a lot of good creamy revenues are not included in that. The 1 trillion number by the way based on our estimate is rising. So we think now it will be an upside to 1 trillion. Plus all these other list of items that I outlisted could be an upside to that for revenues.

[00:26:26]Kjen Sabani from Bloomberg Intelligence.

[00:26:28]Your research is always the go-to man.

[00:26:30]Thanks so much. Let's dive deeper into the Nvidia story, but also like the market reaction. Anna Rafen, founder and CEO of Grenadilla Advisory. You know what? We get ourselves into these frenzies. We're like the Nvidia earnings, they're going to hit and they hit, right? And and you get the red headline on the Bloomberg and it's $91 billion plus or minus 2%. And then everyone goes, but it doesn't even matter what the number is. It's everything else. You're in the markets.

[00:26:58]Why do we behave like that?

[00:27:00]>> It's lonely at the top. I mean, when you're the largest company in the world, I don't think there's anything you can do to surprise investors to the upside.

[00:27:08]So, you become uh a company that needs to defend itself. Um, Nvidia hit it out of the park in many ways. um in almost all the ways that we expected them to not only in the top line and earnings but also you know networking business growing as well as diversifying its customer base everything we've asked for they've given us but I think there are other things like is the growth actually slowing is the growth what we we've been expecting for the last two or three years and that's the question that I don't think Nvidia was able to answer and the dividend increase to me was a big signal that what that the investors had been worrying about might be coming true.

[00:27:49]>> Uh, okay. I actually I apologize to my team in the control room. I'm going to do the opposite of what I just said.

[00:27:54]>> Okay.

[00:27:54]>> Let's linger then. Let's linger on on the shareholder returns. What What were you getting at there? Give me a little bit more.

[00:28:01]>> Yeah. So, you know, the buybacks, it's not that big because if you're, you know, multi- trillion dollar company, 80 billion isn't that huge. I see. It was really the dividend side of it. when a company starts to return dividends.

[00:28:14]That's a multi-year commitment, right?

[00:28:16]So that means while maybe we're not going to be finding other areas to invest in, we're just going to be returning some cash to the investors.

[00:28:24]And that's a signal for growth companies that says, okay, maybe we're not going to be growing as fast as you would anticipated. So that was a small signal.

[00:28:34]We'll see what happens. I still think Nvidia is going to continue to grow because of this massive ecosystem and they're at the center of but it was definitely a signal maybe a a small one but a certain one.

[00:28:47]>> We're showing this image on the screen right and we went into this earnings prior to them hitting saying two things.

[00:28:53]Everyone expects a beat and raise situation but also it doesn't really matter historically what Nvidia says.

[00:29:00]That data I appreciate there's a lot there is how the S&P 500 trades in the days and months that follow and Nvidia earnings even when they beat big and there's no guarantee that the market is going to sort of re respond with euphoria. Um you know does that matter anymore this kind of story that at the index level given Nvidia's size and scale that was kind of what we were all bracing for? Yeah, I mean it matters because of its size, right? The largest company behaving this way is going to have an impact on the index. Now, this reminds me of how how Apple used to behave about 10 plus years ago, there was nothing that Apple could do to satisfy investors. The market would actually punish them for it. So I think this is just what happens when you are at the top and you have high investors have high expectations of you and if you happen to have a high concentration in the index you can take everyone down with you.

[00:29:56]>> It's really really tempting always to look at how a stock trades in the moment and try and find the story. But what I would also say is that you know if you think about last Friday, Monday, Tuesday when the president of the United States returned from China, Nvidia and and the sector, the socks were close to correction territory. You know, we're down again now post earnings on video.

[00:30:17]What what lesson would you learn from that?

[00:30:20]>> You know, there there are two things. I think the the reaction today really is about Nvidia itself. It's not even about the AI industry. It's about Nvidia.

[00:30:29]Friday was a huge disappointment and since then we've learned that China actually banned the gaming chips uh while Jensen Hang was there. So when we're looking at China and Nvidia, it's not a reliable market as much as is the market that we would like to be in. Um so it is I think right for Jensen Huang to sort of pay attention to other areas in which they can expand. Uh think about sovereign AIS outside of China really.

[00:30:57]Um, and I think China would be sort of a bonus, uh, an upside optionality if it happens. But that if is a very big if.

[00:31:06]Anna Rafman, founder and CEO of Grenadilla Advisory. Thank you very much. Now, coming up, we're going to join be joined by Sarah from Conviction to drill down on kind of the impact of Nvidia on a much broader ecosystem of startups. These are companies of all sizes that use the compute or the library of data and and and models themselves. Can't wait. This is Bloomberg Tech.

[00:31:36]Let's get back to Nvidia earnings. A part of Jensen Wong's latest message to investors, Nvidia isn't just selling chips to big tech anymore. It's applying the picks and shovels for an AI gold rush where startups are building everything from AI agents to humanoid robots, robo taxis, everything in between. But on Nvidia compute, joining us, one of the key backers of that ecosystem, Sarah, founder of Conviction, an AI native venture firm. And and I think that's true, right? I think the best place I think to start with you is think about all your portfolio companies. Many of them come on this show and we say, "Okay, what what's the biggest constraint on you right now?

[00:32:12]where you most focus and often times it is compute um and they they're sometimes super proud that they are working on NVIDIA gear um but are they accessing that compute through a cloud provider or are they able to get the cluster themselves direct control >> it depends on what they're trying to do right we have companies that um are you know building at the infrastructure layer building at the model layer building at the application layer one of the first things we did when we started the fund was go buy compute for our companies um just because we can take the timing risk as a venture fund and we concluded exactly what you just described which is they're all going to need it right and um it's >> what did that look like Sarah sorry to interrupt you but like literally >> yeah we we went and bought uh you know H100s at the time a bunch of nodes cloud uh and but the the point was like you know the market has gone through different shortages and now people are predicting a lot of ongoing constrained supply and we were worried about access for startups in that period of time which is why we did Um but our our companies to your actual question if almost all of them want want to do experimentation and they all start with frontier performance which is Nvidia chips today right because that allows you to do new things and then as these companies mature they tend to postrain smaller models think about cost think about how to um be able to uh even just being able to use more tokens on the same task allows you to change the user experience, but everybody wants to start with current generation chips.

[00:33:44]>> The the state of play that Jensen describes is demand vastly outpacing Nvidia, but also in aggregate the industry's ability to supply. On the other side of the table, would you say that that state of play is true for the founders that you're trying to help grow?

[00:34:00]>> Uh, it's probably been a um two quarters of increasing stress in the ecosystem about access to supply at different scales. Wow.

[00:34:10]>> Um, and so I, you know, have spent a good amount of time with the leaders of companies that, you know, serve NVIDIA chips in cloud, um, asking to buy $100 million of a time at of of compute. And I've never been in that scenario before where you're trying very hard to pay somebody a lot of money with a multi-year commitment. Yeah.

[00:34:31]>> Um, and it just speaks to the shortage dynamics. I think it's very hard to get on demand smallcale compute right now.

[00:34:38]um which is what many startups are starting with as as you um um were describing. And so the I think the shortage dynamic is real. And what is more interesting to me is um uh you know Nvidia have this amazing outperformance on the earnings side and Jensen will repeatedly tell people demand is parabolic. humans struggle to consume that. And you know, it's hard to feel bad for Jensen Juan. Jensen's winner in all of this, but you know, he's >> Thank you for being honest.

[00:35:10]>> Yeah. Yeah. But like, and he's an amazing leader. Um uh but he's like, I'm I'm telling you honestly the reality of what I project repeatedly and you just keep not believing me, right? And I I'd say like I believe him from the demand side. Um because what I expect will happen is this you know massive exponential growth that shows up in things like clawed code revenue um is really because you have long horizon agents that people are using in this one use case already. But the world is not just code, right? And so if there's zero, not zero, but small scale, a few billion to 40 billion of revenue in a very small number of months because we figured out how to use the models more productively on long horizon agent tasks. We're going to do that for lots of different functions in the knowledge economy.

[00:35:58]>> Um, while you're here, if you don't mind, like SpaceX S1. Wow. Um, >> yes. Wow. It's inspiring. You have a big event today, right, where I believe like some of the cursor guys will be there.

[00:36:08]And so we we learn a little bit about that. But there's this number total addressable market for AI 26.5 trillion super focused on enterprise. You know, they literally label it as such in the in the deck.

[00:36:24]>> Go ahead. I mean, >> yeah. I I think it's it's a it's a it's a it's a funny number and then like a funny, you know, diagram when you visualize the tan because it's like Starlink and then Enterprise A.

[00:36:34]>> Yeah. Yeah.

[00:36:35]>> Um and you know much of what we invest in is enterprise AI. My my friend Andre Karpathy described this well which is the um the simplest way to think about the use of AI is automation right is automation of tasks we already do right >> um and if we give these models the tools and the harness such that they can do these tasks they can take over things for us. And so I I I do think that the um the opportunity is as big as the other.

[00:37:02]>> Does Elon believe the opportunity or is this him trying to justify why SpaceX took XAI?

[00:37:07]>> I I think it's interesting that you know what you put in the S1 is a choice, right? And he could have easily just said you know like something broader or you know focus on other parts of the SpaceX TAM overall but he is making a commitment that we are going to go make you know uh our offerings relevant in enterprise AI um and so I think he's very committed to this >> yes or no you think he will pull it off >> uh I so there's a question about whether the value is infrastructure models or applications and see this in the you know anthropic deal and the cursor deal as well. Um they have the infrastructure, the infrastructure and the capability to build more is extraordinarily valuable and I think the question will be like I think he's going to make money on that no matter what.

[00:37:54]The question is do they also need to own the model and the application layer.

[00:37:57]>> To Sarah's point, yes, one said that Anthropic would pay SpaceX $1.25 billion per month through 29 for compute. Sarah Gore of conviction across every part of it. Thank you very much. All right, a lot more to come still in the madness of the program. We'll be breaking down two blockbuster IPOs on the docket. Two, SpaceX and Open AI maybe. This is Bloomberg Tech.

[00:38:30]Let's get back to SpaceX. In addition to the company's blockbuster IPO filing, investors are also closely watching tonight's highly anticipated Starship launch. The mission marks SpaceX's 12th major Starship test flight, a critical milestone as the company pushes to advance its nextG rocket program while preparing for what could become the biggest IPO in history. Bloomberg Sana Pashanka joins us now. And you're probably sick of me, right, Sana?

[00:38:55]Because how many times did I post in the blog, Starship is central to all of this. It said it in the S1 could not be more clear. Let's start with tonight's test. What are we expecting? What does it look like? How long does it run for?

[00:39:09]>> Right. So, today, um, you know, Starship is slated to take off from the launch pad at Starbase, Texas at around 6:30 p.m. ET, 5:30 p.m. Texas time. Um, it's going to take off, ignite its Raptor engines, and liftoff into space. Um, it is debuting an upgraded version of the vehicle. It's called version 3 that has upgraded engines. Um, it's supposed to have improved power and capability. Um, it's going to, yeah, lift off into space, lap around the globe, and then eventually splash down into the Indian Ocean. And, you know, as you were mentioning, Ed, this is a really crucial test for this vehicle that is just, you know, so important for all of Elon Musk's ambitions for SpaceX and, you know, the valuation of the company for the coming IPO.

[00:39:56]>> So, let let's talk a little bit about those ambitions, right? The idea is that Starship is capable of carrying a much greater payload to orbit. So think Starlink, think space-based data center and humans getting to Mars. Explain it to us.

[00:40:10]>> Yeah. Yeah. Totally. So um you know Starship has kind of the payload capacity that will enable something let's start with orbital data centers.

[00:40:18]It will enable um an orbital data center of you know as many as 1 million spacecraft orbiting the earth which is what you know SpaceX has said they want to do. um you can't do that to scale as well with the Falcon rockets, their current rockets. You really need a much larger vehicle that can carry 60 satellites to space. Um and so, you know, it will they will need Starship to effectively build these orbital data centers at a timeline and a pace that really makes sense and could actually lead to a profit one day. to get humans to moon to the moon and Mars you'll also need you know that type of capability that type of thrust um Starship is advertised as the most powerful rocket ever built and you know Starship is so important to SpaceX's ambitions but it's also key to America space ambitions because Starship has a contract SpaceX has a contract with NASA for Starship to land humans on the moon >> sign passion card thank you very much fresh off a courtroom in over Elon Musk.

[00:41:22]The race for the trillion dollar AI mega IPO is on. Did you have this on your bingo card? Sources say Open AI may make a confidential filing as soon as this Friday. Joining us now with the details, Bloomberg Sharen Kafari yesterday afternoon. It's the last thing anyone needed was to see that headline hit. But the timing is interesting. What do we know about an OpenAI path to an IPO?

[00:41:47]>> That's right. So, an OpenAI uh filing, it would be a confidential filing, could come as soon as Friday or in the coming weeks. Um, that would obviously put it in uh close competition with Anthropic, which is also expected to potentially file this fall. Um, as well as SpaceX, which is already out there with the S1 listing.

[00:42:06]>> Yeah, I I'm going to let us put SpaceX to one side for a second. So, let's say it's a race between Openropic for September, October. What do we know about about that timeline? But also the numbers involved, right? Do we have a sense of valuation or or what either company wants to raise?

[00:42:21]>> That's right. So the valuations are already, you know, nearing a trillion, right? Open AAI was last valued valued at over 850 billion just uh earlier this spring. Anthropic similarly also um you know in talks to raise at uh up to 900 uh valuation as we've reported. So when we're thinking about an IPO uh months out, you can only, you know, you would hope that for these companies at least they're thinking they want to go further north of that getting us closer to potentially that trillion mark.

[00:42:52]>> There's an idea that when Elon Musk was told by the jury and judge he's not allowed to proceed with his suit, it removed an overhang to open AI, let them kind of proceed with their ambitions.

[00:43:01]But why does open AI want to go public?

[00:43:03]I know it seems silly, but it do they need money? Is it the public perception thing? Has it always been in the plans?

[00:43:09]It's a good question because on the one hand these companies like OpenAI and Anthropic can command these massive private financing rounds that we've never seen before, unprecedented levels of cash. However, access to public markets allows them to do that um with much more ease frequency. It also allows everyday people to get in on the AI boom uh you know or bust which is a risk right if it does change. Uh but that also you know then potentially is allowing a larger group of sort of retail investors to have a stake in the company. really large pools of capital being more easy to access without all the funding round song and dance I think helps them >> shrinking all over the open AI and the entropic story and will be for the rest of this year that does it for this edition of Bloomberg tech this what markets look like not too much going on at the index level basically flat Nvidia is down 1 and a.5% but in the background there's still this idea that the big one is coming SpaceX's IP O this is Bloomberg Tech.