SpaceX Clawed Back Above $150. The Panic Was Over Spending, Not a Broken Business — But the Rally Prices In Perfection

SpaceX hasn't been public for three months, and it has already lived an entire bull-and-bear cycle in that window. The stock (SPCX) opened at $150 on its first day in mid-June, ran to a $192 intraday high that valued the company near $2.5 trillion, then crashed to $104.83 after its first earnings report. Now it's back above $150 and worth roughly $1.95 trillion — up 18% in a month, and about 40% off the bottom.

The easy read of that round trip is that the market panicked and then recovered. The more useful question is what the market was panicking about, because the answer tells you which side of this stock is real.

What scared it was spending, not lost demand

The first shock, on August 4, contained a genuine surprise — but it wasn't on the revenue line. SpaceX beat badly. Revenue came in at $7.81 billion, up 92% from a year ago and about 14% above estimates. Adjusted EBITDA nearly tripled to $3.5 billion. The net loss narrowed from roughly $1 billion a year earlier to $541 million.

The scare was the bill for all of that growth. Capital spending hit $18.4 billion in a single quarter — 235% of that quarter's revenue and 42% above what analysts expected — with roughly 86% of it poured into AI compute. First-half capex ran to $28.5 billion against operating cash flow of just $3.5 billion, leaving about $25 billion of negative free cash flow. The stock fell 13.6% the day after, to a record closing low, and the sell-off spread when as many as 911.5 million shares held by employees and early investors came eligible to trade on August 6.

That is what the panic was about: capital deployment, not customer attrition. The business that the spending is meant to fund kept compounding.

The engine underneath is compounding

The numbers behind the beat are the ones that matter. Starlink — the connectivity business — is the profit engine, and it is widening. It produced $1.66 billion of operating profit on $4.29 billion of revenue, a roughly 39% operating margin, with subscribers doubling year-over-year to 12 million. Connectivity operating income rose 79% year-over-year, and enterprise and government connectivity revenue grew 108% to $1.81 billion, helped by more than $6 billion of new multi-year Starshield contracts.

The AI segment is the uncertainty, but it grew even faster: $2.56 billion in revenue, up 247%, on the back of cloud-compute leases from Google and Anthropic and Grok subscriptions, with another $6.7 billion of cloud contracts already signed for the second half. SpaceX ended June with $47.5 billion of backlog and $100 billion of cash.

This is the classic profile the markets routinely misread: a scale-up spending heavily to build capacity — 1.4 gigawatts of compute, up from 400 megawatts a year earlier — on top of a core that is already throwing off profit. That is a growing-pains story, not a broken-business story. The fear that the capex revealed fundamental damage simply isn't in the operating numbers.

Now the part the bulls don't quote

But the same accounting that makes the case for the recovery also cuts against the current price. The AI segment is frequently described as "EBITDA positive" — a true statement that hides the real math. That figure excludes $1.89 billion of quarterly depreciation on the very machines the company is spending billions to build. Count the depreciation and the segment ran a $1.26 billion operating loss for the quarter.

The other reason to be careful about the AI revenue: its durability is thin. A significant share of it comes from a handful of customers — Google and Anthropic dominate — and the contracts can generally be terminated on 90 days' notice. That is not a backlog like the $47.5 billion of launch commitments; it is a short-dated lease someone can cancel.

And none of this is cheap. At $148 and a roughly $1.95 trillion market cap, SPCX trades at around 85 times trailing sales and more than 350 times EV/EBITDA. It is unprofitable on a GAAP basis, with a negative free-cash-flow margin and negative returns on invested capital. JPMorgan expects capex near $200 billion a year in both 2027 and 2028, and Goldman Sachs does not model positive free cash flow until 2031. For context, Palantir — itself considered an expensive stock — trades near 65 times trailing sales.

About that 2027 price

The headline target floating around — Pivotal Research's $220 for end-2027, with a 12-month median of $216 across dozens of analysts and bulls as high as $800 — tells you the range of what people can justify with a calculator and a straight face. It is not an answer. The same math that produced $220 also produced a $75 low-end target.

What will actually settle the price over the next 15 months is supply as much as demand. The scarcity premium that propped up the IPO float is gone: the August unlock tripled the freely trading shares, the December lock-up expiration will push the tradable float to roughly 58%, and Musk's roughly 42% stake unlocks in June 2027. Every one of those dates is a supply overhang no price target captures.

So the sensible conclusion is narrower and more honest than "buy the dip" or "it hits $220." The market sold SpaceX for the wrong reason — it treated a massive build-out as decay and ignored a core that more than doubled subscribers while expanding an operating margin near 40%. That part was a mistake, and the rebound to $150 reflects a real business, not meme momentum.

But the rebound pricing in near-perfection is the separate risk, and it hasn't been tested. The buyer at $148 is paying more than 50 times forward sales for a company whose AI growth depends on a few customers who can cancel in 90 days, and whose cash isn't expected to turn positive for half a decade. Back above $150 says sentiment recovered. On what's been reported so far, it doesn't yet say the stock is cheap. The single number that would prove the optimistic case — durable, contracted AI revenue from more than two or three short-notice tenants — hasn't been shown yet. Until it is, treat the rebound as the market correcting its earlier misread, not as confirmation that the future is priced right.