Latest / Deep Values: Value Investing Deep Dives / 【IPO】The SpaceX Prospectus: Analysis, Valuation, and Access Guide
Transcript
- 0:00Imagine looking at a corporate balance sheet where a company is
- 0:04just burning through like roughly $2 billion more than it
- 0:08actually makes. Yeah.
- 0:09You'd think that's a disaster. Exactly.
- 0:11Naturally, you'd assume that business is in a full blown
- 0:13financial crisis, right, But Wall Street analysts are looking
- 0:17at that exact same math today and slapping a 1.75 trillion
- 0:22yes, trillion with AT dollar price tag on it.
- 0:26It's. Well, it's honestly
- 0:27unprecedented. It really is.
- 0:28So welcome to the deep dive. Today we are cracking open the
- 0:32massive, highly anticipated Form S1 prospectus for SpaceX, which
- 0:37was officially filed on May 20th, 2026.
- 0:40Finally, right, We are bypassing the hype to really unpack the
- 0:44underlying physics, the frankly wild economics, and the
- 0:47logistics of their target $75 billion fundraise.
- 0:51And, you know, the historical gravity of this filing is just
- 0:53immense. I mean, for nearly a decade,
- 0:55leadership figures like Gwynne Shotwell and Elon Musk fiercely
- 0:58guarded this company. Yeah, they were famously
- 1:00private. Very much so.
- 1:02They actively resisted going public because they wanted to
- 1:06shield their long term multi planetary vision from the, you
- 1:11know, the short term, quarter by quarter pressures of the stock.
- 1:14Market. It's a 90 day earnings cycle.
- 1:16Exactly. So filing this S 1 is a
- 1:18monumental strategic pivot. They're transitioning from this
- 1:22secretive private aerospace shop into a sprawling public
- 1:26conglomerate. And they're expected to start
- 1:28training on the NASDAQ under the ticker STCX around June 12th,
- 1:34right? Yep, June 12th is the target.
- 1:36So to even begin to wrap your head around a $1.75 trillion
- 1:40valuation, we have to kind of throw out your preconceived
- 1:43notions. You really do.
- 1:44If you were listening to this and still thinking of SpaceX as
- 1:47just a rocket company, you are misreading the board entirely.
- 1:50The prospectus lays out something completely different.
- 1:52Yeah, they aren't just building vehicles anymore.
- 1:54The S1 explicitly organizes the company into, well, three
- 1:58distinct segments, revealing them to be this vertically
- 2:01integrated data and logistics monopoly.
- 2:03Which fundamentally redefines their identity.
- 2:05Completely, let's look at the three pillars the South one
- 2:08outlines first you have space. This is the physical logistics
- 2:12side. You know the brute force moving
- 2:13of mass into orbit using the Falcon rockets and start.
- 2:16Right as the classic SpaceX we know, right?
- 2:18Second, you have connectivity, which is the Starlink global
- 2:21satellite Internet infrastructure.
- 2:23And finally, you have the new wild card segment, which is AI
- 2:27driven by their recent merger with XAI.
- 2:29That's the part that threw everyone for a loop.
- 2:32Oh, absolutely. It outlines their massive
- 2:34computational ambitions. Let.
- 2:35Me, try to put this in perspective for you.
- 2:38It's like discovering a local delivery company doesn't just
- 2:42own the delivery trucks. You dig into their paperwork and
- 2:45find out they also paved all the roads.
- 2:48They own the fiber optic Internet cables buried
- 2:50underneath those roads. And Oh yeah, they built the data
- 2:53centers powering the entire city's grid.
- 2:55That is actually a highly accurate way to visualize it,
- 2:59and understanding how those 3 completely disparate divisions
- 3:02connect requires first principles thinking.
- 3:06Meaning you have to strip a business down to the
- 3:08foundational truths of physics and material costs.
- 3:11Exactly. Take the AI segment for example.
- 3:14Why on earth merge a frontier artificial intelligence company
- 3:19with a space logistics company? Right.
- 3:21It sounds like a sci-fi mashup. It does, but it makes sense
- 3:24because AI models require just mind boggling amounts of compute
- 3:29power, and massive compute requires enormous amounts of
- 3:33electricity and, you know, seamless global data routing.
- 3:37O, if you're listening to this and thinking, wait, why is a
- 3:39rocket company suddenly a frontier AI lab?
- 3:42You're having the exact same reaction Wall Street had.
- 3:45Everyone was scratching their heads, Yeah.
- 3:47But when you look at the puzzle pieces, Starling provides the
- 3:51low latency global data routing and the space segment provides
- 3:54the heavy lift delivery mechanism.
- 3:56Because the S1 hints at a future where they deploy off grid
- 4:00orbital data centers. Orbital data centers.
- 4:02Just wow I. Know, and the physics of that
- 4:04are fascinating. Down here on Earth, data centers
- 4:07are severely constrained by the power grid and the massive
- 4:10amounts of water needed for cooling.
- 4:12Right, they use millions of gallons of water.
- 4:14Exactly. But in space you have
- 4:17uninterrupted 247 solar radiation for infinite power.
- 4:21Wow. OK, but what about cooling?
- 4:24Well, cooling is an incredibly complex engineering challenge in
- 4:27a vacuum because you can only cool via radiation, not
- 4:30convection. But the ambient environment is
- 4:33near absolute 0. So they are basically building
- 4:36an ecosystem where the Rockets only exist to deploy the
- 4:40infrastructure that powers the AI.
- 4:43That's the vision. Let's talk about the Rockets for
- 4:45a second, though, because the physics of their economic mode
- 4:48are wild. Let's look at the space segment
- 4:50and what the prospectus calls their reusability deflationary
- 4:54loop. Yeah, that's a key term.
- 4:56I mean, traditional aerospace operated on a single use
- 4:58paradigm for over half a century.
- 5:01They treated rockets like incredibly complex bespoke
- 5:04ammunition, Which. Is such a bizarre concept when
- 5:07you think about it. It's like running a global
- 5:09shipping empire, but instead of bringing your massive cargo
- 5:12ships back to port, you just scuttle them at the bottom of
- 5:14the ocean after a single trip. Right.
- 5:16It's financially ruinous, and it made the cost of getting a
- 5:19single kilogram of mass into orbit astronomically high.
- 5:23Only superpower governments could afford it.
- 5:25So Spacex's core breakthrough wasn't just building a rocket?
- 5:28No. It was booster recycling.
- 5:31By landing and reusing a Falcon 9 or a Starship booster, the
- 5:35marginal cost of a launch just plummets.
- 5:37Because instead of paying for a whole new vehicle, you are just
- 5:39paying for propellant. Exactly.
- 5:41Liquid oxygen and methane, which are, you know, cheap industrial
- 5:44commodities and a bit of refurbishment.
- 5:46By crushing that cost, they essentially manufactured a new
- 5:49market that couldn't afford to exist before.
- 5:52Because space is infinite, but the specific low orbits you need
- 5:55for Internet service are highly constrained.
- 5:58Yeah. Low Earth orbit, or Leo has
- 6:01severe physical limitations. You have collision risks, the
- 6:04threat of orbital debris casting.
- 6:06Tesler syndrome, right? Yes, exactly.
- 6:08And strict international regulatory limits on who gets to
- 6:11broadcast on what radio spectrum.
- 6:13So by driving their launch costs to the floor, SpaceX became
- 6:17their own primary customer. Basically, yeah, They are
- 6:20launching nearly every two days now, now deploying over 7000
- 6:24Starling satellites by 2026. Which is insane volume.
- 6:28It is. They simply got their first and
- 6:30fastest, effectively claiming the prime orbital real estate
- 6:33before competitors could even, you know, build their rockets.
- 6:36OK, so we see the three heads of this Titan space connectivity
- 6:39and AI. Now we really need to follow the
- 6:42money, because looking at the 2025 financials in this
- 6:46prospectus feels like trying to read a foreign language.
- 6:49Oh. It is a massive financial tug of
- 6:51war. It is a company of extreme
- 6:53contradictions. How so?
- 6:55Well, one division is a flawless cash engine and the other two
- 6:58are historic cash drains. I see numbers like this and my
- 7:01eyes just glaze over. So let's break it down for
- 7:02everyone. In 2025, they did $18.76 billion
- 7:07in total revenue, Which? Is massive scale huge?
- 7:10But they posted a consolidated operating loss of $2.59 billion,
- 7:15yet simultaneously they report an Adjusted EBITDA of $6.58
- 7:19billion. Right.
- 7:20How could a company be wildly profitable by one metric and
- 7:23deeply in the red on the exact same accounting document?
- 7:26To understand that, we have to look at capital expenditure or
- 7:29CapEx. Think of it like a neighborhood
- 7:33bakery. OK, I'm trekking the.
- 7:35Daily profit on selling cupcakes, which is basically
- 7:38what EBITDA measures. You know the cash generated by
- 7:40the core operations before the accountant step in is massive.
- 7:44The bakery is incredibly popular.
- 7:46Selling lots of cupcakes. Exactly, but the owners just
- 7:49spent millions building a state-of-the-art 10 story mega
- 7:52oven. That mega oven is CapEx.
- 7:55I see. On paper, their operating income
- 7:57looks negative because they are rapidly writing off the cost of
- 8:00that massive oven. But the underlying cupcake
- 8:03business is brending cash. SpaceX is doing this just with
- 8:07rockets and satellites. And the scale of their mega oven
- 8:09is just staggering. Total 2025 CapEx was $20.74
- 8:14billion. Yeah, that means they spent
- 8:16roughly $2 billion more on infrastructure and development
- 8:20than they even brought in as revenue for the entire year.
- 8:22So. Let's look at the engine funding
- 8:24this. Starlink generated $11.4 billion
- 8:27of that total revenue, producing $4.4 billion in operating
- 8:31income. That is a profit margin of
- 8:33roughly 39%. Wait, I have to stop you there
- 8:36because 39 percent is unheard of for telecom hardware.
- 8:39It really is, but it's because the unit economics don't act
- 8:43like hardware, they act like software.
- 8:45What? Do you mean?
- 8:46Well, the launch business is bottlenecked by physics.
- 8:49Like how fast you can refurbish a rocket, the weather at the
- 8:52launchpad, the literal number of days in a year.
- 8:55Starlink doesn't have those bottlenecks.
- 8:57Oh, because once the satellite is up there, right?
- 8:59Once the fixed cost of launching a satellite is paid, adding a
- 9:02new subscriber at a rural cabin or say, on a cruise ship cost
- 9:06SpaceX almost nothing. It is nearly pure margin.
- 9:09Wow. So Starlink is the cash cow, but
- 9:11then we look at the drains. The space segment brought in
- 9:14about $4 billion, but it is bleeding money because they are
- 9:17pouring capital into developing Starship.
- 9:19Yeah. And then there's the AI segment.
- 9:21The AI division is the true black hole on the balance sheet.
- 9:24How bad is it? It posted $3.2 billion in
- 9:28revenue, but a catastrophic $6.4 billion operating loss.
- 9:32Of that total $20.74 billion in CapEx we just discussed, a full
- 9:3761% of it went entirely to the AI segment.
- 9:40Unbelievable. And I'm looking at this line
- 9:42item for the space division's R&D and I just can't get past
- 9:46this number. They spent $15 billion
- 9:49developing Starship between 2020 and 2025.
- 9:52Yeah, it's a huge number. I don't.
- 9:53Care how you spend it, spending $15 billion on experimental
- 9:56rockets right before you ask Wall Street for money feels
- 9:59reckless. How do the underwriters even
- 10:02stomach that? Well, to a traditional
- 10:04accountant it looks chaotic, sure.
- 10:06But to an aerospace historian, it looks like the greatest
- 10:10bargain in human history. A.
- 10:11Bargain 15 billion. Yes, because we need a baseline
- 10:15to understand that. 15 billion to build the Apollo program.
- 10:19NASA spent roughly $280 billion in inflation adjusted money.
- 10:23SpaceX achieved roughly a 95% reduction in development costs.
- 10:27Wait. Really.
- 10:27A 95% discount on a moon rocket? How is that physically possible?
- 10:31Radical vertical integration and rapid iterative prototyping.
- 10:34Breaking things fast. Exactly.
- 10:37Traditional aerospace spends years running computer
- 10:40simulations to ensure the rocket works perfectly on the first
- 10:43try. SpaceX builds hardware
- 10:45incredibly fast, tests it on the launchpad until it explodes,
- 10:49analyzes the debris, and builds the next one that.
- 10:52Sounds expensive but I guess it saves time.
- 10:54Huge amounts of time. Furthermore, they aggressively
- 10:57control material costs. Instead of using fragile,
- 11:00bespoke carbon fiber that requires massive clean rooms,
- 11:04they pivoted to 300 series stainless steel.
- 11:06Wait, isn't that the same cheap alloy used in water?
- 11:09Towers it is, And paradoxically, stainless steel actually gets
- 11:12stronger at the freezing, cryogenic temperatures of liquid
- 11:15oxygen in space. It's heavier, but it's
- 11:17drastically cheaper and faster Manufacturer.
- 11:20OK, so we accept they are incredibly efficient engineers
- 11:22and Starlink is highly profitable, but they are still
- 11:25burning billions overall largely due to that AI compute drain,
- 11:29right? So if Wall Street banks are
- 11:31projecting a trillion dollar payday based on a company
- 11:33burning cash, how are underwriters like Goldman Sachs
- 11:36and Morgan Stanley justifying a $1.75 trillion valuation?
- 11:41Well, it's a math problem that demands absolute perfection.
- 11:45The prospectus reveals that even if Space XS revenue and EBITDA
- 11:48were to completely double next year in 2026, which is a.
- 11:52Huge if. Right.
- 11:53A 1.75 trillion valuation still implies a multiple of 56 times
- 11:59forward revenue and 109 times forward EBITDA.
- 12:03Just for context for you listening, a high flying tech
- 12:05company like Tesla trades at roughly 12 times forward
- 12:08revenue. Palantir is around 43 times
- 12:11revenue. Yeah, there are simply no true
- 12:14public comparables for a business that is simultaneously
- 12:17a highly profitable telecom, a dominant aerospace manufacturer
- 12:21and a loss making Frontier AI Lab.
- 12:23So what Wall Street is essentially doing is asking you
- 12:25to pay today for massive, incredibly successful future
- 12:28businesses that simply do not exist at scale yet.
- 12:31Exactly. You are buying a vision of the
- 12:33future, not present cash flow. The underwriters model this out
- 12:37in 3 scenarios. Break those down for us.
- 12:38OK, so in a bear case, a Starlink user growth hits a
- 12:40ceiling, Starship development stalls, and the AI monetization
- 12:44fails to keep up with the massive of hardware costs.
- 12:46You're looking at a company worth maybe 500 to $800 billion.
- 12:50Which is a terrifying drop from 1.75 trillion.
- 12:53What does a base case look like? A base case assumes Starlink
- 12:56continues its steady global expansion, Falcon rockets
- 13:00maintain their market monopolies, Starship becomes a
- 13:02usable asset, and AI improves but remains incredibly capital
- 13:06heavy. I.
- 13:07Have a valuation there. That puts the valuation
- 13:09somewhere between 900 billion and $1.3 trillion.
- 13:13So even in a highly successful dominant base case, 1.75
- 13:17trillion is still massively overpriced.
- 13:20Yes, that $1.75 trillion number sits firmly at the upper limit
- 13:25of a bull case scenario. Got it?
- 13:26To justify paying that price today, you have to believe
- 13:29Starlink becomes a ubiquitous high margin global utility.
- 13:32You have to believe Starship Florida State materializes to
- 13:35unlock a totally new orbital economy.
- 13:38And crucially, you have to believe that the AI segment
- 13:41translates its massive compute drain into durable high margin
- 13:44revenue. I noticed the company itself
- 13:46claims a total addressable market or Tam of $28.5 trillion,
- 13:50but reading between the lines of the S1, that mostly seems tied
- 13:53to their future AI ambitions and those off grid space data
- 13:56centers. Yeah, that Tam is a statement of
- 13:59extreme vision, not a modelable cash flow for next year.
- 14:02Starlink alone, even globally dominant, cannot support a $1.75
- 14:07trillion price tag. It requires flawless future
- 14:11execution across all three segments.
- 14:13OK, so let's assume you're listening to this, you've
- 14:14absorbed the risks, but you believe in the bull case you
- 14:17want in, right? How does an everyday person
- 14:20actually buy into an IPO of this magnitude before it hits the
- 14:24open market? Usually, Wall Street banks
- 14:26aren't just handing out primary shares to retail investors.
- 14:29They usually don't. In a hot IPO, institute
- 14:32institutional investors, you know the massive hedge funds and
- 14:35mutual funds, consume 90 to 95% of the shares at the offering
- 14:39price. Leaving crumbs for everyone
- 14:41else. Exactly.
- 14:42Everyday retail investors get maybe 5 to 10% if they're lucky.
- 14:45But Project Apex, which is the internal name for this 21 bank
- 14:49syndicate led by Goldman Sachs and Morgan Stanley, features a
- 14:52highly unusual anomaly. What?
- 14:54Kind of anomaly. The reporting indicates Elon
- 14:56Musk is demanding up to 30% of this IPO be allocated directly
- 14:59to retail investors 30. Percent.
- 15:01That's huge, yeah. They are even hosting 1500
- 15:05global retail investors at a major pre IPO event.
- 15:08Which dramatically shifts how the stock will trade on day one.
- 15:12But let's talk actionable mechanics.
- 15:14If you want those shares, you can't just assume your everyday
- 15:17phone app will ping you with an invite, no?
- 15:19Definitely not you. Have to navigate the brokerages
- 15:22based on the specific banking syndicate.
- 15:23So where should people look? The highest probability route
- 15:26for a standard retail investor is through E*Trade because
- 15:30Morgan Stanley is a lead underwriter.
- 15:32Their retail subsidy subsidiary, E*Trade, is actively expected to
- 15:36handle the smaller ticket retail flow.
- 15:38So E*Trade is your highest confidence Ave. yes?
- 15:41Bank of America, Citi and UBS are also in the syndicate, but
- 15:45their allocations appear heavily biased toward domestic high net
- 15:47worth clients, family offices and international distribution.
- 15:51What about the massive consumer apps?
- 15:53Robin Hood. Sofi.
- 15:54Fidelity. Schwab.
- 15:56You need to be incredibly careful here.
- 15:59Well, Robin Hood has an IPO access feature and the others
- 16:02have similar tools. They're currently unconfirmed.
- 16:05The reporting does not show them officially in the SpaceX selling
- 16:08group yet. It is possible they get an
- 16:11allocation as the syndicate broadens, but right now you
- 16:14cannot bank your strategy on them.
- 16:16Good to know. Are there any other ways?
- 16:19Well, if you missed the primary allocation, there is one
- 16:21indirect backdoor route. Platforms like Sofi and Titan
- 16:25offer access to the ARK Venture Fund, ticker ARKVX.
- 16:30Right, Kathy Woods Fund. Exactly.
- 16:32This is an actively managed fund holding private venture back
- 16:35stakes including private shares of SpaceX.
- 16:38It allows a backdoor method to own a piece of the company prior
- 16:41to the public float though you are buying a blended fund, not
- 16:45pure SpaceX. Before anyone rushes to move
- 16:47their capital though, we really have to talk about the
- 16:49governance risk because the fine print here is thick.
- 16:52Explain what the dual class share structure actually means
- 16:55for a regular person buying the stock.
- 16:57It means this is not a democratic shareholder
- 16:59structure. The S1 reveals Musk is retaining
- 17:01roughly 85.1% of the combined voting power through super
- 17:05voting shares. So you are essentially a
- 17:07passenger. You get the financial upside,
- 17:10but you have zero hands on the steering wheel.
- 17:12Your ability to influence corporate direction is
- 17:15practically nonexistent. Exactly 0 Furthermore, the
- 17:19prospectus notes that roughly 1/5 of their 2025 revenue came
- 17:23directly from the federal government, primarily NASA and
- 17:27defense contracts that. Feels like a double edged sword.
- 17:30It is. You are buying into a company
- 17:32that operates on the bleeding edge of physics, but also on the
- 17:36bleeding edge of federal regulation and geopolitical
- 17:39risk. Which means if political winds
- 17:41shift, or if regulatory agencies clamp down on orbital space
- 17:45debris, a massive chunk of the reliable revenue could just
- 17:48vanish overnight. Yes.
- 17:50Which brings us to the ultimate long term risk.
- 17:52Right. Which leaves you with one final
- 17:54provocative thought to Mull over Pull directly from the risk
- 17:57factors deeply buried in the prospectus, SpaceX themselves
- 18:01explicitly warned that their most ambitious future projects,
- 18:04specifically those orbital AI data centers and deep space
- 18:07infrastructure, may never become commercially viable.
- 18:10It is the ultimate question you must answer for yourself before
- 18:13June 12th. I mean, right now the AI segment
- 18:16is burning 61 percent of a company's capital expenditure.
- 18:21They're competing against the infinite balance sheets of big
- 18:24tech, the Microsoft's, Googles and Amazons of the world.
- 18:28If that AI segment fails to scale, if those orbital data
- 18:31centers succumb to the harsh physics of space and never
- 18:34materialize, what happens to that $1.75 trillion sci-fi
- 18:38valuation? Exactly, it's a massive bet and
- 18:41to access more advanced researches, applications and our
- 18:44latest published personal intelligent Multi agent Deep
- 18:47value intelligence. Welcome to explore and join Deep
- 18:50Values dot AI community. And if you want to track total
- 18:53market temperature, welcome to search and download the iOS app
- 18:56Buffett Indicator from the Apple App Store.
- 18:58Disclaimer All analysis is based on publicly available sources
- 19:01and may be subject to revisions or differing interpretations.
- 19:04All contents from this channel is for information and research
- 19:07purposes only and does not constitute investment advice.
- 19:09Conduct your own due diligence before making investment
- 19:12decisions. If you like to access wider
- 19:14topics including life and investing philosophy, please
- 19:16subscribe to our sister channel Deep Values.
- 19:19Don't forget to like this video. Much appreciated.