Loading summary
Transcript56 lines
- [00:00]
A
Foreign. Welcome to the AI to roi, the Big Story podcast. I'm Ray Reich, your host or co Host, the CEO of BenchMarket. And joining me as always, is my co host, Peter Buchanan, Managing Partner of New Plan. Peter, we're covering the SpaceX IPO today and I want to lead with a spoiler.
- [00:29]
B
Well, just go for it, just eliminate the suspense.
- [00:33]
A
Well, SpaceX's launch business is genuinely great. It's heavenly. Starlink is really great and the IPO as currently structured, not so great. So, hey, we're not here to give investment advice. That's not what our show does or our background. We're here to present data and facts about what's going on in the AI and AI adjacent industries and tell you what it actually says.
- [01:04]
B
That's exactly right. SpaceX, they filed for a confidential IPO registration with the SEC on April 1st. The public S1 is expected at the end of the month or beginning of May with a NASDAQ listing targeting early June. And it would be the largest IPO in history by a factor of 2.5x at a valuation of $1.75 trillion. They're raising $75 billion. At a valuation of $1.75trillion. There's never been an IPO this big.
- [01:43]
A
Oh, the next closest one actually was Aramco. Right, the Saudi oil company, and it raised 25 billion. I still remember that.
- [01:52]
B
Yes, at a 4x valuation.
- [01:54]
A
Exactly. Okay, well, let's give you some of the inside baseball details. So first, SpaceX merged with XAI in February of this year. And what it really did was it took a high cash burn. Some people might call it a bleeding AI business based upon two operations because you had XAI and you had X and investors are now being asked to pay almost 95 times 20, 25 revenue for that combined entity.
- [02:27]
B
Right. And so today we're going to walk through the business piece by piece. That's Launch services, Starlink, xai X. We're going to talk about the IPO mechanics and we're going to talk about the post IPO structure and what it actually looks like because there's some changes to the rules for this ipo. So let's start with what works.
- [02:52]
A
Well, let's start with Launch services then, because it's the real, I'll call it a monopoly. And by the way, it's also really cool eye candy when you see those rockets blasting off or, my gosh, them coming back down and being caught right on the way down, it's pretty amazing.
- [03:07]
B
And you know, so you know the son of a friend of ours, Ray, actually designed the nose cone of that rocket.
- [03:14]
A
I do. My former GE connection, I believe.
- [03:17]
B
Yeah, that's right. Yeah.
- [03:18]
A
Well, so Launch Services is the part of SpaceX that everyone actually respects and pretty amazed by it. Right. And those Falcon launch costs, like a Falcon 9 launch costs about $67 million. And if you compare it to the competition, like the United Launch alliance, it runs 110, 160 million. So it's very efficient. Efficient for a rocket launch service. And by the way, there's no other meaningful competitor at scale. And we heard about Jeff Bezos and Blue Origins, New Glenn, right. And they just launched commercially for the first time last year. You got Rocket Labs, Neutron and. Right, they, they suffered a tank failure earlier this year and won't fly at least until Q4 at the earliest of this year. So they are far ahead of the competition. They hold about $4 billion in NASA contracts and they just won a almost 180 million Space Force contract in April. So they are the 800 pound gorilla. And rocket Launch services, they are.
- [04:21]
B
And that price advantage alone is really extraordinary. I mean it's 40 to 60% cheaper for an incredibly reliable service. I mean they do well over 100 launches a year. So that's a reliability record that no one's going to match for the foreseeable future. So what's the catch?
- [04:42]
A
Well, one of the things I actually love about Elon Musk is he's truly a visionary and he communicates as a optimistic, visionary entrepreneur. So Starship, which is their next generation rocket, it was initially supposed to enter commercial service in 2021. Now it's five years later and it's completed 11 test flights. And the next launch is scheduled for mid May with a new version, their V3. So one of the things I would say is sometimes Elon's a little bit more optimistic than physics and reality can actually support. And even their own internal document documents put their first orbital refueling so be able to refuel, you know, in space in June 2026. And they're not going to have an uncrewed lunar landing, so moon landing until June 2027. So they still got a lot of heavy lifting in front of them, even though they're by far the leader.
- [05:44]
B
Right. And Starship isn't optional. The next generation Starlink satellites require it. The orbital data centers that Elon talks about pretty endlessly also require it. The NASA Artemis crew moon landing which is targeted for 2028 requires it. So if Starship stays behind schedule, everything else in the bull case, their bull investment case lives with it.
- [06:13]
A
Now there's a second catch, and I see both the cautionary tale here, but also the positive, and that is the size. You know, we all talk about target addressable market. Well, the global space launch market, you know, it reached about $18 billion in 2025 and it's projected to grow to 32 billion by 2032. So over the next seven years, that's only a 14.5% CAGR. So it's a solid business. But hey, at 32 billion total market, even if he gets 75% of it, probably doesn't justify a 1.75 trillion valuation. In fact, to justify it, SpaceX would have to be priced higher on a multiple basis than almost every hyperscaler and valuable tech company in the world. So the launch monopoly earns a premium, but I think the premium might be a little high. But there's more to the story and we're going to talk more about that AI story in a little bit.
- [07:15]
B
Right. But it really all comes back to starship. The orbital data centers that Elon describes the economics that they want to get to for those, it's just brutal. And so TechCrunch basically says, related to these orbital data centers, that the cost of operating them is basically 3x the cost of operating a ground based data center. And we don't really know the economics at all of that. Elon just says we're going to do it and it's going to be great. Now he said the same thing about the launch services business when he started it, and nobody thought he could get it to scale. He has gotten it to scale and he's sort of done it on time and he has this monopoly. But orbital data centers, that was incredibly hard. This just seems much harder. Let's go to Starlink, because Starlink is also a space miracle.
- [08:17]
A
Yeah, it just doesn't get enough press coverage because Starlink is really impressive. They have 10 million subscribers around the world as of February. That's almost $10 billion in revenue from 2025. Now that's about 54% of SpaceX's total revenue. So this is a business that's been subsidizing a lot of the capital intensive, you know, launch services. And it's interesting to see the revenue mix because it really does matter here. You know, residential subscribers like me, we pay 120 to $150 a month. Right. But what people don't understand is residential is not the key. It's the other verticals, like maritime. Maritime Those ships out there right in the Hermu Strait, you know, they're using Starlink and they pay around $34,000 in annual revenue per customer. Then we got aviation. I mean, my gosh, I've been waiting for a year for United finally to launch their Starlink based WI fi on the airplanes. The aviation average is about $300,000 annually per customer. And it's those top two tiers, aviation and maritime, that generates the margins that make the rest of the unit economics work.
- [09:34]
B
Right. They've got some headwinds, but they're definitely manageable. So Amazon rebranded its Project Kuiper satellite efforts as Amazon Leo and they've launched an Enterprise Beta. But they only have a little less than 250 satellites in orbit versus Starlink's over 10,000. Amazon has a deadline to deploy over 3,000 satellites in a constellation by July of this year. So far they only have 700. They think they'll have 700 by that date. So they're asking for an extension. They're going back to the teacher and saying, please, can I have a little bit more time. So there's not much really getting in Starlink's way.
- [10:18]
A
I tell you, Peter, I got to tell you this. I was listening to the FCC commissioner today on CNBC. SpaceX, Starlink has a application in for a million, a million satellites. So, right, you talk about the 10,000 today. That tells you about the grandeur of Elon's vision here. But part of the challenge is market access like Starlink still hasn't launched in India after years of negotiation. Africa, which could really benefit from this service, is untapped. And on the consumer side, you know, SpaceX, honestly I think I got mine initially for $50 a month. They've been cutting residential prices to compete with the wired providers, right, the FiOS from Verizon, et cetera. So I think their residential business is going to face increased margin compression challenges. But what about the mobile side?
- [11:18]
B
Yeah, so SpaceX, they also want to, they want to launch a Starlink mobile service. We can use our cell phones anywhere in the world, independent of tower access. And so they went and bought for about $19 billion, like 8.5 billion in cash, 8.5 billion in stock and some interest from EchoStar. They bought EchoStar's spectrum. So EchoStar after many years decided they didn't want to be in the mobile business. Then they went and did a partnership with T Mobile to embed that capability inside a phone sold by T Mobile. And that service that they want to launch, they Want it to be an incredibly high speed service beyond telephone calls. And so it requires very high speed 5G data and it requires the new Starlink V2 satellites, which requires the starship to deploy them. So again, that starship is really important. And so the timeline for getting that really going is 2027, about the time the new 6G standard starts to come out. So they have a challenge there.
- [12:31]
A
You know, Peter, I'm still, I'm very bullish on SpaceX launch services and Starlink, but no matter how good of a business executive you are, when you have a multi business unit strategy, or heck, for me as a parent. Right. You might always have that problem child. Right. So that problem child for Elon right now in this IPO is xai, because it's complicated. You want me to give you some details?
- [13:01]
B
Sure.
- [13:02]
A
So. So SpaceX acquired XAI in February of 2026 in an all stock deal valued at $250 billion. So think about that. XAI worth 250, and you're thinking about a 1.75 valuation for the IPO. Now that transaction also included X. X became part of Xai, which was formerly Twitter, which Musk acquired in 2022. And at that point in time the valuation was reported to be 1.25. That's the SpaceX Starlink XAI and X altogether $1.25 trillion.
- [13:40]
B
Right. And then some financial analysis came out. So just a few days ago, the information reported that if Xai had been part of SpaceX in 2025, the combined entity would have posted a net loss of $5 billion on $18.5 billion in revenue. Prior to the merger in 2024, SpaceX had generated $8 billion in profit on $14.2 billion in revenue. So that's a dramatic reversal as a result of this transaction.
- [14:18]
A
It is. But at the same time, I think about what the hyperscalers are doing and investing in their AI infrastructure. I mean, they're consuming almost all of their free cash flow margins. Right. And we're talking 50 billion, 100 billion. Amazon said 200 billion. So I'm not too struck by the magnitude, but it is a huge risk. And let me give you some of the Xai financials specifically. So Xai, which is their competitor to chat GPT, OpenAI, to Google, Gemini and to Anthropic, they burned through almost $9.5 billion of cash in nine months, the first nine months of 2025, with only 210 million in revenue. So that's a very high burn rate, like burning $28 million of cash today, their capital expenditures for Xai neared 13 billion for 2025. So that's, you know, what is that like more than 50% of.
- [15:15]
B
It's over a billion dollars a month.
- [15:17]
A
Exactly. And they are carrying $12.5 billion in debt. So there was a lot of risk there. But I also stand investing ahead of the major AI revenue trend. But then there's the reality of Grok, which is their large language model. Where do they stand competitively today? Peter?
- [15:36]
B
Right, well, before we get to Grok, I just want to say for all those hyperscalers spending all that money, even after spending all that money, they're still profitable. Yes, they're still profitable. So they might be spending all that free cash flow, but they're not bleeding the cash that they're raising. Right. So basically, let's get to Grok here. So GROK is basically behind, behind, behind, behind. So an internal SpaceX memo described Xai as clearly behind its competitors, Anthropic, OpenAI, Gemini, the leading Chinese models. And recently Elon Musk said publicly that GROK wasn't built right and needed to be rebuilt. Most of the enterprise customers they have are actually customers where Musk has existing relationships already. Companies like Morgan Stanley and Palantir and they're generating off those customers on average per customer, single digit millions in revenues. They don't have very many eight figure customers yet that, you know, OpenAI and Anthropic and Google certainly do. So they've made some progress though.
- [16:53]
A
Yeah, well they did Cross Grok crossed $1 billion in annualized consumer subscription revenue. So that's pretty impressive, right? Getting to $1 billion. And I think that was announced at the Xia, all hands before the merger. But they did invest 9.5 billion to get there. So there is still that time lag and gap between what they're generating and what they're burning. And you know, if they were kind of a leader in the industry, I wouldn't be as concerned. But man, the market leaders aren't slowing down. What's going on? And Anthropic chat GPT, even at Meta, they are moving quickly too, and they're
- [17:34]
B
growing much, much faster. I mean, we talk in the newsletter about Anthropic's incredible growth rate and we talked about it last week on this podcast. They've gone from 9 billion in ARR to 30 billion in ARR in a couple months. There's no growth like that here. And now let's turn ourselves XAI is another challenge because they have Incredible leadership turnover. The CFO is leaving this month and it's part of a wave of senior exits. In fact, Elon is the only one of the 12 XAI founders who's still with the company in advance of the ipo. Elon is actually ordering XAI to cut spending on anything that doesn't generate revenue.
- [18:29]
A
I actually like that. That sounds like fiscal discipline. But you know what? Elon's always been great and famous for firing people. I still remember when he fired the majority of their charging network team at Tesla. I drove, you know, 800 miles this weekend and I will tell you, it sure didn't hurt my experience as a Tesla supercharged network customer. So I'm not going to get too up in arms about that. And the other thing I would say about Grok, right, it's getting trained partially on the data from X, formerly Twitter. So that gives us some real time insight into social discourse that other models can't replicate. So you know, if I was said, hey, train on Twitter, ie X or Reddit, I think they have a competitive advantage there. Not that that necessarily is directly correlated to additional revenue generation. I think over time it's going to help.
- [19:23]
B
Right. All right, let's move to the IPO mechanics because they're pretty interesting. So several things stand out. So we mentioned before, SpaceX is raising $75 billion at a target valuation of 1.75 to $2 trillion. 95 times revenue at 95 times revenue going out in June. Samantha Lau, who's the chief investment officer at Alliance Bernstein, told the Information, this is a game of chicken and sentiment in this market in the last two years. Everything is about the narrative, not the fundamentals.
- [20:03]
A
Right? And let's admit Elon Musk does have a halo, a business halo. And in fact, we talked about that in the newsletter this week. And there are some things that are interesting about these IPO mechanics. Number one, the banker selection. Right. Five banks are leading the offering, including B of A, Citigroup, Goldman Sachs, JP Morgan and Morgan Stanley. Notice I listed those alphabetically because no single bank is holding that traditional lead position. Right. And at the same time, mustard require each bank to purchase a Grox subscription or subscriptions as a condition of participation. Now, I haven't seen the exact details of that and I guess you could call that a lock in, but hey, the virtuous cycle of AI investing in customers hasn't seemed to bother most people. But I heard one customer actually agreed to spend over $10 million on their Grok subscriptions.
- [20:58]
B
Right? And then there's a retail allocation that's three times the typical size. So 30% of the available shares are going to go to individual investors. So if you're a fan of Elon, you can actually probably get this. You have a better chance of getting this stock. And then there's a NASDAQ rule change that is the most controversial element of the whole structure. Ray, explain that.
- [21:24]
A
Yeah. They recently announced that for their index inclusion changes on May 1, traditionally, companies ranking the top 40, the NASDAQ 100 by full market cap. Full market cap can now join the index after 15 trading days. That's down from 90 days. And the reason that's so important is it can create a structural buying mandate for those passive index funds like Invesco's QQQ over 363 billion assets. So you're going to see what could be portrayed as an artificial accelerated buying of that stock 15 days after they're listed by these mega funds.
- [22:11]
B
Right. And so people who deal with these markets all the time are extremely unhappy about that. So George Noble, who's a 45 year veteran of asset management, calls it the most shameless structural manipulation of a major index he'd ever seen. Patrick Healy of the Issuer Network told Bloomberg that anything shorter than 30 days is unnecessarily risky. And Owen Lamont of Acadian Asset Management raised a core issue, which is index funds are supposed to accept the prices the market gives them, but not buy before a reliable price is established. And after 15 days, there's still a lot of hype there. It hasn't settled down.
- [22:54]
A
That is going to introduce more risk to the retail investor. But also the institutional investors seem to be a little cautious about this. Right. The information reported that some have told the bankers that a $2 trillion valuation would be unpalatable. So good thing it's priced at 1.75 trillion right now. So I guess that's a little bit south of unpalatable. But the reality is, and this happens a lot in capital intensive businesses, SpaceX needs the IPO proceeds, right? That Echo Star Spectrum acquisition we talked about, I think they have $2 billion in debt to help facilitate that. Right? They need to help, they need to help fund that. The Starship development is more capital intensive than expected. And come on, the Orbital Data Center Colossus Data center project needs additional capital.
- [23:46]
B
It does. And by the way, you know, Xai, before the transaction in January, they raised $20 billion at a $250 billion valuation. And without, without the fundraising that they just spending it on. Xai they would be out of money by next summer based on their current spending profile. So they really need the proceeds from the ipo.
- [24:13]
A
We're going to need to wrap up here. And as a additional caution to our listeners, we are not an investment advice podcast. But I have a question. What's the bull case in your mind? And what's the bear case in your mind, Peter?
- [24:31]
B
Okay, so the bull case is basically these things have to happen. Starship reaches sustainable commercial operations within 18 months. So it's lifting off a lot. It's delivering those version 3 satellites. It enables the mobile service Groq starts to generate measurable enterprise revenue growth beyond Elon's friends. The PASIG index fund buying mandated by NASDAQ provides a floor that supports them for a while. Frankly, Elon does have a track record of delivering things, right? Just think of some of these accomplishments.
- [25:15]
A
Come on. Falcon 9 reusability, who would have thunk, right? Starlink, one of the fastest growing $10 billion companies you've probably never heard of, right? And then there's Tesla's EV position. Who would have thought 15 years ago that electric vehicles would be as strong and Tesla would be leading that not just in the United States but overall. So he has a track record of big wins. He does have the track record of missed timelines like Starship being five years late. And even the robo taxi and some of the full self driving Teslas, they are a little late. But I will tell you Peter, on the bullish case, I think all these overly optimistic changing the world and reality as we know it, that may be Elon Musk feature, not a bug.
- [26:04]
B
Well, he has it. He's going to have it the rest of his life because it's just what he does. Now the bear case is pretty bearish. So if Starship encounters further delays, that undermines the growth of launch services and of Starlink. If XAI can't get its costs under control, it's still burning 28 $30 million a day. That's very bad. They've really got to have a sales engine to sup up some of those expenses. And once the lockup periods expire, there's going to be significant selling pressure.
- [26:43]
A
Yeah, and in fact the economists actually described this combined entity as Musk. Bet that includes the fates of Xai, SpaceX and Tesla, that they're going to be increasingly intertwined. So that framing is fairly accurate. Things like the Falcon 9 launches Starlink satellite at marginal cost. Starlink provides a global network infrastructure to distribute GROK XAI models, improve Starlink's networks management orbital data centers can give SpaceX a structural AI mode that no terrestrial competitor can replicate. Think about all the issues around water, electricity, etc. Here's what I think. I think it's a genuine vertical integration strategy. It goes beyond a mash up of somewhat related companies to a vertical integration strategy. And by the way, he pulled that off at Tesla, right? I mean he built the manufacturing facilities, he built the chips. You know, he started on Nvidia, then he built his own AI chips. So I don't count Elon out. And I still remember where you and I met. We talk about this once in a while. We met at GE and Jack Welch positioned these 11 very different business units as components of the same growth engine. And even though I never got that because the business units really didn't collaborate and they weren't vertical integration supply chain partners, I think in this case we do have a vertically integrated supply chain that could pull it all together.
- [28:25]
B
Right. But every element in that strategy has a dependency. There's the Starship dependency, there's the economics of the orbital data center, there's the GROK capability map capability gap. And their sales engine, there's X has to stop losing advertisers right now. X's number of monthly active users is now below Threads and Threads is only a two year old service. So you can have these dependencies. There's a downside to dependencies.
- [28:58]
A
So yeah, but you know what, even though without Xai, boy, if this was just SpaceX launch and Starlink, it would be a fantastic IPO with long term legs. That launch monopoly is real. Starlink's market position is strong and by stapling XI onto this, it's an impressive operational and financial success story. And if it becomes successful, and honestly, I think Elon is so committed to XAI in X because of the history. I mean everyone on the podcast hopefully remembers he was a co founder of OpenAI with Sam Altman in 2017. He left because he did not think that open AI should be a for profit, generate lots of wealth for their shareholders. So I'm pulling for them. But what do you think the executives listening to our podcast should take away from this? Peter?
- [29:59]
B
I think there's three things. So first the valuation multiples. They're not valuing fundamentals, they're basically asking you to bet on things that are going to happen. The big bang moment for this company is in the early 2000s if they succeed. I agree. Right. I think that's really the first thing. So when you think about vendor pricing, vendor stability. If they're signing multi year contracts and they get, basically, if they can sign multi year contracts, if they can get Starship working, they will grow into this valuation, but they won't be doing it for three or four years.
- [30:41]
A
But when you've got the federal government and the Department of Defense, as I won't call it a backstop, but as one of their largest customers, that's a good place to start from. And you know, cyclically, I think the synergy between SpaceX and XAI is coherent. Right. Starlink as a global AI distribution infrastructure, training grok on X's real time data, orbital data centers as a structural moat. It's an interesting story. The reality of that story, you know, some people saying 18 to 36 months out. Maybe it's 36 months to 60 months out, but it's an interesting vision.
- [31:16]
B
Oh for sure. And then this NASDAQ rule change is worth watching beyond SpaceX because this fast entry mechanism is going to help AI companies like OpenAI and Anthropic and there are a few others that are queued up that are not as blingy, that are going to go public perhaps before OpenAI and they're going to benefit going out on NASDAQ from that rule change as well. But let me also say again, we are not providing financial advice in this podcast we are analyzing the state of these incredibly interesting space and AI company called SpaceX.
- [31:58]
A
Peter, take us home. What's your summary here?
- [32:00]
B
So here's where we read. I look at the SpaceX IPO. So the first, the two businesses that work Launch Services and Starlink are attached to one that absolutely doesn't. And then it's priced as if all three are operating at full potential simultaneously and somehow much bigger than they actually are. And the evidence just, there's nothing in the fact set that says that a $1.75 trillion valuation is, is any way justified.
- [32:35]
A
Yeah. And you know, here we are, AI to roi. So we need to be driven by the numbers, not just the passion and vision. And right now the numbers don't say this is a 1.5, $1.75 trillion IPO. But the market will tell us.
- [32:56]
B
It will tell us. And so if you find this podcast useful, well, we have the AI to ROI newsletter, the link is in the show notes, but we'll tell it to you anyway. It's AI, the number two roi.substack.com AI to roi.substack.com we hope you'll sign up and read us five days a week
- [33:21]
A
and all the details that we shared while you're there on the treadmill or driving listening to the podcast. They are there in the Monday, April 13th edition of the AI Roi Newsletter. Thanks, Peter. This was fun.
- [33:35]
B
It was.