Elon Musk Warns SpaceX Stock Short Sellers: Can Musk’s $100 Trillion SPCX Bet Work?

On the night of July 17, a stranger asked Elon Musk on X why short sellers kept betting against SpaceX. Musk's reply was one sentence: anyone holding a serious short position in SpaceX for long has, in his words, a "very low" chance of surviving. It read like a threat dressed up as a fact.

Here's what Musk didn't mention in that reply. At the exact moment he typed it, SpaceX's stock, ticker SPCX on the Nasdaq, was trading at its lowest level since its record-breaking IPO five weeks earlier. Shares had fallen nearly 47% from their all-time high. Short sellers, by multiple independent estimates, were sitting on billions of dollars in paper profits. If this is a fight between Musk's conviction and the market's current verdict, the market is winning right now.

Let's break down what Musk's warning actually means, where the eye-popping $100 trillion vision behind it comes from, and everything that has to go right, gate by gate, in rockets, in cash flow, in physics itself, for that vision to hold up. Then we'll check what the numbers say about who is actually winning the short-seller fight today, and what that leaves you with at current prices.

The Warning Landed on a Stock That Was Already Bleeding: Why Is SpaceX Stock Falling After Its IPO?

Musk wasn't replying to a journalist or an analyst. He was replying to a random X user, in an exchange that also touched a much bigger claim. In a separate post the same week, Musk told space entrepreneur and early SpaceX investor Peter Diamandis that SpaceX would eventually be "worth more than Earth" if the company hits its goals. Big talk, even by Musk's standards. The stock chart tells a very different short-term story.

Sources: Google Finance

Do the math on that and you get a stock down close to 47% from its high, and about 11% below the price at which it first sold shares to the public. In IPO language, that makes SPCX what traders call a "broken IPO," one that has fallen below its own offer price. It's an unusual place for the biggest listing in stock market history to find itself, just five weeks after Wall Street called it a record breaker.

None of this means SpaceX the company is in trouble. Revenue is real, Starlink is profitable, and the rockets keep flying (mostly). But it does mean Musk's warning landed at a moment when, so far, the trade is going the short sellers' way. And underneath that single tweet are really two separate questions that get conflated constantly, including by Musk himself:

Can shorting SpaceX be dangerous? And is SpaceX reasonably valued at its current price? The honest answer is that the first can be true, short selling a volatile, thinly floated stock with a huge retail following genuinely is risky, while the second stays completely open. Musk being right about the risk to shorts says nothing about whether the stock is cheap. The rest of this piece treats those as the two separate questions they are.

Where This "$100 Trillion" Number Actually Comes From

This is worth clarifying up front: no bank on Wall Street has published research saying SpaceX is worth $100 trillion. That number comes from Peter Diamandis, the XPRIZE founder, who argued around the time of the IPO that SpaceX wouldn't just become a $10 trillion company someday. It could become the first $100 trillion company in history. His logic: all the wealth humans have built on Earth, real estate, gold, every stock market combined, adds up to somewhere around $600 trillion. Space, in his framing, holds resources that make even that number look small. Musk's own reply to Diamandis, that SpaceX could end up worth more than Earth itself, is really an endorsement of this same idea, not a separate claim with its own math behind it.

Compare that to what SpaceX has told its own regulators. In its IPO filing, the company estimated its total addressable market at $28.5 trillion, dominated by the AI opportunity, a genuinely enormous number, but a fraction of Diamandis's $100 trillion. Wall Street banks covering the stock, including UBS and JPMorgan, have cited similarly sized figures, in the neighborhood of $30 trillion.

Here's the trap in treating any of these TAM figures as a stock price target: a total addressable market is not revenue, and revenue is not shareholder value. Think of a new expressway connecting two cities. The highway can unlock hotels, warehouses, factories, and logistics parks worth thousands of crores along its route. But the company that built and operates the toll road doesn't own all of that economic activity. It owns the toll booth. 

SpaceX is trying to be several things at once here, the toll road, the truck manufacturer, and in some markets the business actually using the road, and how much value it captures depends entirely on which role it plays in each new industry.

A cleaner way to think it through is to pass any TAM number through five filters before believing it's relevant to a share price.

Run a number through that funnel and it shrinks fast. A hypothetical $10 trillion future industry might generate only $500 billion a year in actual revenue once real customers and real pricing show up. If SpaceX captures 10% of that, it's earning $50 billion. At a healthy 30% operating margin, that's $15 billion of operating profit, a real number, but a rounding error against the original headline. This is exactly why SpaceX's own $28.5 trillion figure, and certainly Diamandis's $100 trillion, describe an opportunity SpaceX wants to chase, not a forecast of what it will earn.

How Proven Is SpaceX's Business Today? Revenue, Losses and Cash Burn Explained

Before getting into what still has to happen, it's worth being precise about what has already happened:

Sources: SpaceX S-1 filing, SEC EDGAR, reporting via Yahoo Finance, SatNews, and Hargreaves Lansdown, May-June 2026.

Two things jump out. First, revenue grew a genuine 33% in 2025, and Starlink alone is a real, profitable, fast-growing business, not a story stock waiting to happen. Second, the AI segment is now the dominant use of cash. It burned $12.7 billion in 2025, up from $5.6 billion in 2024 and just $463 million in 2023, and then burned $7.7 billion in the first quarter of 2026 alone, more than its entire 2024 spend in one quarter. Total 2025 capital expenditure across the company ran close to $20.7 billion against roughly $6.8 billion of cash generated from operations, a gap plugged with new debt and equity. 

The 2024 comparison matters too: before the AI segment was folded into SpaceX's books via the xAI merger, the company posted a $791 million net profit for that year. The AI bet is what turned a profitable rocket-and-internet company into one still posting billion-dollar quarterly losses.

None of this makes SpaceX a bad business. It does mean investors buying SPCX today are not simply buying today's SpaceX. They're pre-paying for several future versions of it that don't exist yet.

The Eight-Gate Ladder: What Must Starship Achieve for SpaceX’s Valuation to Hold?

Nearly every exotic part of the bull case, orbital data centers, beamed solar power, asteroid mining, Mars, depends on one machine working the way Musk says it will. So it's worth asking the question in the most literal way possible: not "will Starship succeed," but how many separate, sequential things have to go right before Starship creates enough value to justify any of this.

Across its full development history, Starship has flown roughly a dozen integrated test flights, with a rough tally of about seven counted as some form of success and the rest ending in failures or partial failures, a track record that looks perfectly normal for an experimental rocket program and completely inadequate for an airline-style business, which is exactly the distinction that matters here. A prototype aircraft completing one test flight doesn't make an airline. An airline needs hundreds of predictable departures, trained crews, spare parts, and paying passengers showing up every day. Starship needs the equivalent before any of gates six through eight become real.

The two most recent flights make the point directly. In May 2026, the newest version of the rocket, Starship V3, suffered an engine failure on its upper stage during its first flight, and several booster engines failed to relight during the return burn, a setback serious enough to ground the program for 52 days while the FAA investigated. The very next flight, on July 16, 2026, was aborted on the pad itself after four of the Super Heavy booster's 33 engines failed to ignite. That flight was meant to carry Starship's first real cargo, 20 Starlink satellites, and test an in-space engine restart that NASA needs demonstrated before trusting the rocket with astronauts. It's since been rescheduled, with SpaceX targeting July 23. Orbital refueling, gate six, hasn't been attempted at all yet. NASA's own inspector general estimated in a March 2026 report that the Artemis moon-landing program needs somewhere between 10 and 16 successful propellant-transfer flights before a crewed landing becomes realistic, with that landing currently targeted for early 2028.

History offers a blunt warning about assuming reusability alone solves this. The Space Shuttle was also designed to make spaceflight routine and cheap through reuse. In practice, refurbishment between flights stayed complicated and expensive, launch frequency never came close to what was promised at the program's start, and costs remained stubbornly high for its entire operational life. SpaceX has a better design philosophy and a real track record of cutting costs with Falcon 9, but Starship's economics still need to be judged by measured turnaround time and refurbishment cost, not by whether a single vehicle lands upright.

There's a second, less obvious catch buried in the cost story. Musk's argument requires the cost of putting a kilogram into orbit to fall by something like 90%. But "cost per kilogram" has two different meanings that don't have to move together: what it costs SpaceX internally, and what SpaceX charges the customer. If SpaceX cuts its own costs by 90% while remaining the dominant launch provider, and Falcon 9's history suggests SpaceX's internal cost reductions have historically run well ahead of the price cuts it actually passed on to customers, it can simply pocket much of that gap as margin instead of passing it to the market. That would be excellent for SpaceX's profitability and considerably less powerful for creating the new industries the $100 trillion thesis depends on, because those industries only get built if outside companies actually receive cheap enough launch prices to make their own businesses work.

Now stack all eight gates together mathematically, and the compounding effect becomes the single most important thing to understand about this entire vision. Say, purely as an illustration, that each of the eight gates independently has a generous 80% chance of going SpaceX's way. Multiply eight 80% probabilities together "80% x 80% x 80% x 80% x 80% x 80% x 80% x 80%" and the combined odds of clearing all eight fall to roughly 17%, not 80%. Push the assumption to a highly optimistic 90% per gate, and multiplying it out eight times still only gets you to about 43%. These are illustrations, not a forecast of SpaceX's actual odds, and the gates aren't fully independent in real life; success at one probably raises the odds at the next. But the lesson holds regardless of the exact inputs: a bet that depends on many separate breakthroughs should never be priced as though every one of them is a foregone conclusion.

Can Starlink Support SpaceX’s Current Valuation?

Starlink is the one part of this story that isn't a projection. As of the end of Q1 2026, the network had 10.3 million subscribers across 164 countries, had deployed more than 9,600 satellites, and generated $11.39 billion in 2025 revenue at a 63% adjusted EBITDA margin, according to SpaceX's own S-1 filing. It is a real, profitable business carrying the rest of the company.

It still faces four structural constraints worth understanding before assuming the lead is permanent.

Satellites don't last forever. Low-Earth-orbit satellites have useful lifespans generally estimated around four to six years, which means Starlink behaves less like a telecom tower that runs for decades and more like a commercial vehicle fleet that needs constant replacement, even once subscriber growth slows down. Cheaper Starship launches help with this math, but the company still has to keep manufacturing thousands of replacement satellites indefinitely just to stand still.

Orbital traffic is getting more crowded, which adds regulatory and collision-avoidance overhead as more constellations launch, a cost that shows up in operating complexity rather than the income statement.

Competition is real, just further behind than the headlines suggest. Amazon's satellite broadband unit, renamed Amazon Leo, is licensed by the FCC for a 3,236-satellite constellation, with 1,618 of them, half the total, required in orbit by July 30, 2026. As of mid-2026, Amazon had roughly 270 to 400 satellites actually deployed, meaning it will almost certainly miss its own deadline by a wide margin; the company has formally asked the FCC for a 24-month extension. Europe's answer, the IRIS2 network, doesn't have a single satellite in orbit yet and isn't expected to begin service before 2029. Starlink's lead is not just real, it currently looks closer to unassailable than contested.

The number that ultimately matters isn't subscriber count anyway, it's revenue and profit per customer after accounting for subsidized hardware, support costs, and the ongoing replacement cycle above. A much larger but less profitable subscriber base wouldn't support the same valuation as a smaller, genuinely profitable one, and that distinction has gotten less attention than the headline subscriber figure deserves.

The Exotic Bets: Can SpaceX Build Data Centers, Solar Power and Mining Businesses in Space?

This is the part of the $100 trillion vision that depends entirely on the eight gates above actually clearing first, and even then runs into its own separate, unresolved problems.

Orbital data centers sound simple: put AI computing hardware in orbit, run it on constant sunlight, skip the land and grid constraints of Earth-bound data centers. The physics are considerably less forgiving. A data center converts much of its electricity into heat, and on Earth that heat is carried away by air and water. Space is a vacuum. There's no air to carry anything away, which means every watt of computing power needs to be matched with large physical radiators to reject heat, adding weight, complexity, and failure points to launch, assemble, and maintain. On top of the thermal problem sits data movement: shuttling the enormous volumes of information that large AI training runs require is far harder between orbit and Earth than between racks in the same building, which means the more realistic near-term use for orbital compute is processing data that's already generated in space, like satellite imagery or intelligence workloads, rather than training the world's largest AI models. There is real progress here, a Nvidia-chip-equipped satellite has already demonstrated in-orbit AI processing and drawn serious venture funding, but it's a proof of concept for a narrow use case, not evidence that hyperscale training in orbit is close.

Space-based solar power is the most scientifically validated of the exotic bets, and still the least commercially proven. A Caltech-led experiment has already demonstrated detectable wireless power transmission from orbit, so this isn't purely theoretical anymore. But the amount of power involved in that demonstration was tiny next to what a real electricity grid needs, and the economics remain genuinely uncertain: some academic models suggest it could eventually become cost-competitive if launch and manufacturing costs fall enough, while other analyses, including NASA-linked reviews, question whether it can ever beat the pace at which terrestrial solar, batteries, and grid technology keep getting cheaper on their own. Starship might make the space side of this equation better. It doesn't freeze the ground-based side in place while it does.

Asteroid mining has the most eye-catching theoretical numbers and the least real business behind them. Asteroids may hold water, iron, nickel, and platinum-group metals worth a fortune at today's prices, but that math falls apart the moment you account for supply: if a company actually returned a large quantity of platinum to Earth, the additional supply would push the price down, meaning the asteroid can't be valued as if every kilogram sells at the price that exists before you've mined it. There's an even sharper irony buried here. The most likely first commercial product from asteroid mining is water, split into propellant for spacecraft already in orbit, not exotic metals shipped home. But if Starship succeeds at making launches from Earth extremely cheap, as the whole thesis requires, water launched from Earth becomes a cheaper competitor to water mined from an asteroid. Starship's own success could undercut the economics of one of the earliest businesses it's supposed to enable. Add an incomplete international legal framework around who actually owns extracted space resources, and this stays a genuine long-term frontier rather than a dependable source of near-term cash flow.

Why Mars Should Not Be Included in SpaceX’s Current Valuation

Mars is central to SpaceX's mission and almost irrelevant to a responsible stock valuation, and it's worth being direct about why.

Building even a small permanent settlement requires transportation, energy, housing, food production, radiation shielding, and continuous resupply chains that don't exist yet. Terraforming the planet, turning it into somewhere with a breathable atmosphere and livable temperatures, runs into a limit that has nothing to do with SpaceX's execution and everything to do with chemistry. NASA-funded research published in Nature Astronomy concluded there simply isn't enough accessible carbon dioxide on Mars to warm the planet using any technology available today. The study's lead author put it bluntly, calling the idea "not possible using present-day technology." Musk has publicly disagreed with that conclusion, but it remains the mainstream scientific position, and even optimistic engineering proposals point to timelines measured in centuries, not investment cycles.

A project that might generate revenue several generations from now has essentially no present value in a normal valuation model unless you assume both a very high probability of success and an enormous future payoff, an assumption nobody can responsibly make today. Mars belongs in this story as a mission that attracts talent and capital, and as long-term strategic optionality worth acknowledging. It shouldn't be used to justify a meaningful part of what you pay for SPCX right now.

Are SpaceX Short Sellers Winning After the SPCX Stock Crash?

Musk's confidence about short sellers isn't coming from nowhere. He's said versions of this before, and history has occasionally backed him up, at least on Tesla.

In 2018, he warned Tesla short sellers they had "about three weeks" before their position exploded, a boast that arrived the same year as his infamous "funding secured" tweet, which triggered an SEC settlement, a $20 million fine, and his temporary removal as Tesla's chairman. In 2020, Tesla stock surged more than sevenfold, and according to S3 Partners data, short sellers lost more than $40 billion that single year, the worst-performing short trade S3 had tracked at the time. In 2023, Tesla shorts lost another $12.2 billion, more than any other single US stock that year. Add it up, and S3's Ihor Dusaniwsky has put Tesla's cumulative short-seller losses since the company went public at $61.8 billion. As he told CNN Business, "that's going to leave a mark." It isn't a one-way street, though. In 2022, when Tesla stock fell 65%, short sellers actually made $15.9 billion betting against it, as per the same S3 data. Shorts aren't always wrong. Sometimes they're early. Sometimes they're simply right.

Which brings us to what's actually happening with SPCX, rather than what happened with Tesla years ago. Short interest here hasn't just been high, it's been on a near-vertical climb since the IPO.

SPCX's tradable float is approximately 646 million shares. Figures represent different reporting dates and may use slightly different methodologies; treat the trend, not any single number, as the takeaway.

That climb has been extremely profitable for the people on the short side. Bloomberg reported short sellers had booked around $3.88 billion in collective paper profit by July 15; by the next day, July 16, that figure had reportedly jumped to around $8.7 billion, adding nearly $5 billion in a single trading session as the stock kept sliding and the short position kept growing. Even Michael Burry, the investor who made his name shorting the 2008 housing bubble, reportedly looked at shorting SpaceX and passed.

Two things pushed short interest into this range so quickly. SpaceX disclosed plans for a new senior unsecured debt offering, on top of the $75 billion it had just raised in equity, and it separately announced the roughly $60 billion, largely stock-funded acquisition of Cursor's parent company. Both read to some investors as a company financing itself under pressure rather than from a position of strength.

Here's where Musk's framing gets more complicated, not less. When short interest was still modest, back in late June, S3 Partners itself described the setup as looking more like ordinary price discovery than a classic, supply-constrained squeeze, since shares were still easy and cheap to borrow. That characterization is harder to make now. With short interest above 28% of a float this small, some traders have started describing the position as "a lot of potential fuel" for a squeeze if a strong catalyst, a clean Starship flight or a good earnings surprise, forces a wave of short covering all at once. Both readings can be true at different points in time: this may have started as normal, healthy skepticism working itself out in the price, and it may now be building into something that could snap violently in either direction depending on what happens next. Musk's warning might eventually be proven right, the way it eventually was on Tesla. It just isn't proven yet, and the shorts have been the ones collecting so far.

One more mechanical event sits directly in this path. Somewhere between roughly 900 million and 1.37 billion additional shares become eligible to trade in the weeks after SpaceX's Q2 earnings, the gap reflecting two separate lock-up tranches, an unconditional release of about 911.5 million shares roughly two trading days after earnings, plus a further 455.8 million shares that only unlock if SPCX is trading above $175.50, well above today's price. If insiders sell into that unlock while the stock is still this cheap, it adds real supply right as short covering could be adding real demand, a genuinely two-sided setup that nobody can call with confidence in advance.

SpaceX Stock Price Targets: What Wall Street Analysts Expect

Once SpaceX's IPO underwriters cleared their quiet period in early July, a wave of formal analyst coverage arrived, and it's about as divided as coverage gets.

Sources: Named analyst notes as reported by TipRanks, Benzinga, StockAnalysis.com, and Yahoo Finance/The Motley Fool, June to July 2026. Consensus targets across data providers cluster between roughly $212 and $244.

A $63 to $800 spread on the same stock, covered by many of the same major banks, tells you something. This isn't a case of a few analysts being sloppy, it's a case of genuinely different views on which of the eight gates above turn green. Raymond James's $800 target effectively assumes Starship succeeds at a scale that makes SpaceX worth more than $10 trillion. Morningstar's $63 estimate assumes the AI segment stays a drag and most of the exotic upside never shows up in cash flow. Both firms have access to the same public filings. They just disagree, sharply, on which future is more likely.

The current price of $119.85 sits almost exactly on top of CFRA's Sell target of $115. That's either a sign the market has already priced in most of the bad news CFRA was worried about, or a sign there's further to fall if its underlying thesis keeps playing out. Both readings are defensible. Neither is a prediction.

Doing the Reverse Math: Is SPCX Stock Overvalued?

Rather than debate whether $119.85 is cheap or expensive in the abstract, it helps to work backward from the price and ask what it actually requires you to believe.

At a market capitalization of roughly $1.6 trillion, SPCX currently trades at approximately 85.5 times 2025 revenue and around 243 times 2025 adjusted EBITDA. Those are simplified market-cap multiples, not full enterprise-value multiples, and they don't adjust for SpaceX's sizeable debt and preferred stock. But they show plainly how much future growth is already baked into the price, even after a 47% correction.

Now run it forward instead of just looking at the multiple today. Assume, purely for illustration, that an investor wants a 10% annual return over the next nine years. For today's roughly $1.6 trillion value to compound at that rate, SpaceX would need to be worth about $3.77 trillion by 2035.

This isn't a price target or a full discounted cash-flow model, and it excludes interim cash flows, future share issuance, and the additional capital SpaceX will likely need to raise along the way. It's an expectations test. SpaceX generated $18.67 billion in revenue in 2025. Getting to $359 billion by 2035 requires revenue to grow more than 19 times over. Getting to $538 billion requires nearly 29 times growth. Neither number is achievable through launch services and Starlink alone; it requires several of the exotic bets above becoming real, profitable businesses on top of an already-large connectivity franchise.

A scenario table makes the sensitivity easier to see than any single number.

Illustrative model discounting a hypothetical 2035 value back to 2026 at a 10% annual rate; not a forecast.

At roughly $1.6 trillion today, SPCX is priced closer to the exceptional-execution scenario than the conservative one. The 47% fall from the June peak reduced the amount of optimism baked into the stock. It didn't remove it.

The Bear Case: What Are the Biggest Risks for SpaceX Stock?

Put the pieces above together and the risks are not subtle. Starship has had two rocky flights in three months, an engine failure and a 52-day grounding in May, an aborted launch in July, and the entire exotic side of the bull case depends on this specific rocket working reliably and cheaply. The AI segment, recently rebranded SpaceXAI, is burning cash faster than any other part of the business, its capex nearly tripled from 2024 to 2025 and then nearly matched the entire prior year's spend in a single quarter of 2026, with no confirmed timeline to profitability. Short interest above 28% of a genuinely small float means volatility in either direction is now, to a real degree, mechanically manufactured by the short book itself rather than by fundamentals alone. And somewhere between 900 million and 1.37 billion shares becoming freely tradable in the weeks after the next earnings report is a real supply event that has nothing to do with how well the business performs.

None of this makes SpaceX a bad company. It does mean several of the pillars holding up the $100 trillion number are either unproven, underfunded relative to their ambition, or, in Mars's case, currently understood to be out of scientific reach. It's also worth remembering that SPCX isn't the first hyped, story-driven mega-IPO to pop hard and then crack just as hard. That pattern, a huge first-week rally on retail excitement and forced index buying, followed by a sharp reset once the lock-up and the first real earnings report force the market to price the business instead of the story, shows up almost every time a company goes public on narrative before it has a long track record. It doesn't tell you where SPCX goes next. It does tell you that a stock's first two or three months of public trading are usually a poor guide to what it's actually worth.

Our View: A Framework

We're not going to tell you SPCX is a buy, a sell, or a hold, and nobody being straight with you should claim to know that with confidence five weeks into a stock's public life. What we can offer is a way to organize everything above so you can form your own view.

If you believe Starship gets past its current run of engine problems within the next couple of years, that Starlink keeps its commanding lead over Amazon and Europe's satellite networks, and that the AI business finds a path to profitability after the Cursor deal closes, then even SpaceX's own, relatively conservative $28.5 trillion market estimate could support a much larger company than today's price implies. That's a real, coherent bull case sitting in the "execution growth" row above. It just hasn't graduated to "proven" yet.

If instead you think the last two Starship flights are an early warning sign rather than noise, that an AI unit which is now the company's single largest cash drain needs more than one large acquisition to fix itself, and that thirty-one percent short interest plus a nine-figure lock-up landing right after the company's first-ever earnings report is a recipe for real two-way volatility rather than a one-way squeeze, then there's a reasonable case that today's price, close to CFRA's bearish $115 target, still has room to move toward Morningstar's $63 estimate before it moves toward Raymond James's $800 one, if it ever does.

Both are honest readings of the same set of facts. What tips the scale, for anyone weighing this, are two dates on the calendar rather than anything Musk says on X: the Q2 earnings report, SpaceX's first as a public company, and the days immediately after it, when the market finds out whether insiders actually sell into the unlock or choose to hold. Whatever conclusion you reach about SPCX before those events is, by definition, a bet on Musk's framing over the numbers. Whatever conclusion you reach after them is a bet on what the numbers and the insiders actually showed.

Diamandis might turn out to be right that this becomes the first $100 trillion company. NASA's own scientists might turn out to be right that at least one pillar of that vision, a warmer, wetter Mars, simply isn't happening with any technology we currently have. Both of those can't fully be true at once, and the stock price five weeks after IPO doesn't resolve which one wins. Eight gates still stand between here and there, and right now, only one of them is showing green.