Elon Musk Says SpaceX Short-Sellers Probably Won’t Survive. These 3 Earnings Numbers Say He’s Wrong.
Last month, Elon Musk took to X to tell the short-sellers betting against his rocket company, Space Exploration Technologies (SPCX +9.65%), that their “survival probability” is “very low.”
And after its first quarterly report was released last week, a whole lot of investors think he’s right. The company gave bulls plenty to like, with revenue up 92% year over year (YoY), the doubling of Starlink subscriptions, and adjusted earnings before interest, taxes, depreciation, and amortization (EBITDA) nearly tripling.
But dig a little deeper into the numbers, and I think there are plenty of reasons to think the short-sellers — who profit when share prices fall — are right.
Space Exploration Technologies
Today’s Change
(9.65%) $12.86
Current Price
$146.15
Key Data Points
Market Cap
Day’s Range
$134.01 – $149.60
52wk Range
$104.83 – $225.64
Volume
1.7M
Avg Vol
125.9M
SpaceX’s cash burn is staggering
SpaceX generated $3.5 billion in operating cash flow through the first six months of 2026 — impressive, until you see that the company made $28.5 billion in capital expenditures (capex).
Free cash flow (FCF) was roughly negative $25 billion in just six months. And that actually understates the economic investment somewhat, because another roughly $3.9 billion of capital expenditures were financed rather than paid in cash.
Now, to be fair, this is a company in the middle of an enormous build-out, and heavy spending today can mean big payoffs down the road — in theory. Spending at this pace and on this scale puts a whole lot of pressure on those investments to pay off quickly, and, at least for the spending on AI, there’s a real question of whether they will pay off at all.
One customer accounts for nearly 20% of revenue
One unnamed customer accounted for 19.5% of SpaceX’s entire Q2 revenue and the lion’s share of its AI revenue. That could be a serious problem if the customer backs out, especially given that AI is where SpaceX is spending lavishly — $23.6 billion of its total capex went into AI alone.
Image source: Getty Images.
Although the customer is unnamed, it is very likely that it’s Anthropic, the maker of Claude. The deal to lease computing capacity from SpaceX’s xAI is cancellable by either side with just 90 days’ notice after an initial ramping period. That is not the kind of agreement you want when you’re committing billions to service it.
We’ve already seen how loose these arrangements can be. Musk himself previously clarified that one heavily touted lease was initially just for 180 days, despite the potential for a much longer relationship.
Starlink’s hidden problem: falling revenue per user
Starlink is still SpaceX’s best business, generating $4.3 billion of Q2 revenue and $1.7 billion in operating income.
But I think investors have been ignoring a problem: Average revenue per user (ARPU) has been falling. The figure came in at just $66 per month, and though that’s little changed from the previous quarter, it’s down from $85 a year earlier.
ARPU decline is fine when you have subscriber growth to compensate — and make no mistake, Starlink very much does at this point — but as time goes on and more of the market is captured, especially in more developed nations, I think ARPU could start sliding even faster even as growth slows.
Is SpaceX stock overvalued?
Now, I’m not advocating that you short SpaceX stock — that’s a very dangerous maneuver that can easily backfire if you’re wrong — but I am saying that short-sellers will ultimately be proven right. I believe that SpaceX stock is overvalued and that the top-line growth is overshadowing some serious flaws beneath the surface.
And, remember, during the next year, nearly $6 billion of shares owned by early SpaceX investors and employees will be unlocked and available for sale on public markets. Even a sliver of those insiders deciding to sell could put pressure on the stock price.
