SpaceX Earnings Beat, Stock Dropped. The Disconnect Is Bigger Than the Headlines

SpaceX beat its first earnings report as a public company, and the stock fell. Revenue came in at $7.8 billion, well ahead of the $6.8 billion consensus. The net loss narrowed to $541 million from an expected $2.1 billion. Loss per share was 9 cents against a 24-cent forecast. And shares tumbled 8.8% in after-hours trading.

I'm not here to explain away the decline. I'm here to figure out whether the market is reacting to a genuine problem or scaring itself on the wrong variable.

The variable that spooked investors was capital expenditure. SpaceX spent $18.4 billion on capex in the second quarter — roughly $16 billion of it on AI infrastructure, including data centers and orbital compute efforts. That topped the $13 billion analyst estimate. Annualized, capex runs toward $45 billion. That number is enormous for a company generating $7.8 billion in quarterly revenue. It's also the reason the stock's free cash flow margin sits at -144.8%. The company burns far more on long-term investment than it generates from operations.

The question isn't whether that's a lot of money. The question is whether the market just priced SpaceX like a value trap because the capex line item scared it, even though the underlying growth engine — Starlink — is accelerating and the AI business is closer to scale than the headline loss suggests.

Let's break down what actually happened.

Starlink is the machine that funds the rest.

Connectivity — dominated by Starlink — was the only profitable segment. Revenue in the segment jumped 66% quarter-over-quarter. Starlink subscribers reached 12 million, slightly below the 12.19 million estimate, but the miss was marginal and the growth trajectory is clear. The unit generated positive adjusted EBITDA in a business that still requires massive satellite deployment. That's the engine. It's the reason SpaceX can spend $16 billion on AI without running into a liquidity wall.

The balance sheet supports the spending. SpaceX holds $93.5 billion in cash and cash equivalents with total debt of $65.6 billion, leaving net debt at negative $60.7 billion. (Net debt is total debt minus cash — a negative number means the company is net cash-positive.) The quick ratio — which measures whether a company can cover its short-term obligations with its most liquid assets — sits at 499%. This is not a company at risk of running out of money. This is a company choosing to spend on growth.

AI losses beat expectations, and the contracts are real.

The AI segment lost $1.26 billion on operations, roughly half the $2.39 billion analysts feared. Revenue is there — AI segment revenue grew 247% year-over-year and hit roughly $2 billion in the quarter, beating the $2.08 billion estimate. The contracts are real: $920 million per month from Alphabet (Google) through mid-2029, and a similar agreement with Anthropic PBC. Those aren't speculative bets. They're multi-year compute commitments from companies that don't sign deals on impulse.

Musk acknowledged on the call that the AI unit requires "sustained investment before it can generate profits consistently." That's the understatement of the decade. But the beat on both revenue and the operating loss shows the unit is closer to inflection than the consensus was pricing in.

The valuation gap is real — but it's not the gap the market is looking at.

Here's where things get interesting. SpaceX trades at 76 times trailing sales. That is not a cheap number. It's a premium price for a company that hasn't produced GAAP earnings. The EV/EBITDA multiple sits at 320x. By any traditional valuation metric, the stock is expensive.

But the forward growth rate is also in a different universe. Revenue grew 92% year-over-year. The quarter-over-quarter revenue growth was 66.5%. Gross margin stands at 51.9%, which is the kind of operating leverage you see in software and network businesses, not rocket companies. The market has arguably baked in a narrative that the capex burn will continue indefinitely without return. If Starlink scales and AI compute revenue compounds at even a fraction of that growth trajectory, the current multiples compress rapidly because the denominator — earnings — grows faster than the price.

The disconnect I'm watching isn't whether the valuation is cheap today. It's whether a 92%-growth, 52%-gross-margin company with $93 billion in cash gets punished because its capex line item frightens investors who are used to seeing profits, not buildouts.

The moat is still intact.

Before calling this a buying opportunity, the competitive moat needs to survive the scrutiny. On the launch side, Falcon 9 remains the workhorse of the global launch market, with robust cadence. SpaceX just won a $1.6 billion Space Force contract in late July and national-security satellite deals worth roughly $6.5 billion earlier this year. Those aren't discretionary contracts — they're sticky government relationships.

Starlink's moat is orbital infrastructure. There are 4,000+ satellites in low-Earth orbit forming a constellation no competitor can replicate without years of lead time and billions in capital. ARPU has declined as SpaceX enters international markets and rolls out lower-priced plans, but subscriber growth and coverage expansion offset that for now.

On the AI side, the moat is less proven but structurally interesting. SpaceX is building orbital data centers — satellites equipped with NVIDIA chips for compute in space. That's a play no hyperscaler can replicate. The partnerships with Anthropic, Google, and NVIDIA aren't just revenue; they're validation that the orbital compute concept has institutional believers.

Starship remains the wild card. The 400-foot rocket has burned $15 billion to date and hasn't reached commercial service. A successful July 24 test demonstrated larger V3 satellite deployment. Musk claimed Starship flight cadence could reach one launch per day within a year, citing a "very robust" heat shield. That's Musk-speak, which means take it with a grain of salt. But Starship is the flywheel that connects all three segments — cheaper satellite deployment for Starlink, infrastructure for orbital AI, and next-gen launch revenue.

Price action: the post-lockup bounce

The stock has fallen 53% from its $225.64 peak in mid-June, hitting a low of $104.83. The decline was driven by the post-IPO hangover, broader tech sector cooling, and then the earnings-related capex scare. But here's what's happened since the first lock-up expiry on August 6th, which freed 911.5 million shares worth roughly $101 billion for potential sale.

Instead of collapsing, the stock bounced. As of today's close, SPCX is at $133.11, up 15.8% on the day and 22.8% over the past five days. Volume ran 242 million shares with $30.6 billion in turnover. That's not a death spiral. That's absorption.

The RSI sits at 52.8, neutral territory — not oversold, not overbought. The MACD is still negative at -8.0, meaning the longer-term momentum trend hasn't fully flipped bullish yet. But the volume-backed bounce after the lock-up expiry suggests the initial wave of selling has been met with buyers willing to step in. There are 16 tranches of lock-up releases scheduled through December 2026, so the selling pressure isn't over. But the market didn't break on the first test.

The bear case deserves a sentence.

Elon Musk predicted revenue could hit $1 trillion by 2030, with a "non-zero chance" of 2029. FactSet's 2029 estimate is roughly $207 billion — a factor of five difference. The capex burn at $43.3 billion trailing twelve months is unsustainable as a fraction of revenue unless growth accelerates even further. The AI segment is cash-negative and the space segment is cash-negative. If Starship fails to reach commercial service, the entire orbital compute thesis loses its deployment vehicle. And the staggered lock-up schedule means selling pressure will arrive in waves through year-end.

Those are real risks. None of them are priced out of the stock — the 53% decline from peak already reflects significant pessimism. But the bear case is about execution and timing, not about the core business model being broken.

What it all means

The market reacted to capex, not to the business. Revenue beat. Losses narrowed. Starlink is printing. AI contracts are multi-year and binding. Cash on the balance sheet covers several quarters of the current burn rate. The stock fell 53% and bounced 16% in a single day after the lock-up test.

This isn't a stock I'd chase at $133 after a 16% single-day move. The better risk/reward sits on a pullback toward the $115-$120 zone, or after price action confirms the bounce isn't a dead-cat recovery. The lock-up schedule through December means volatility will continue. There's room for more selling.

But the fundamental gates are passing. Revenue growth of 92% year-over-year with 52% gross margins and $93 billion in cash isn't a company whose thesis has broken. It's a company whose buildout phase is expensive, messy, and currently unpopular.

SPCX is a Buy on weakness. I'd reassess if Starlink subscriber growth stalls below 10% year-over-year, if the AI operating loss widens beyond $2 billion per quarter without corresponding revenue growth, or if the stock breaks below $100 and holds there — that would signal the lock-up selling has more legs than the current bounce suggests.

For now, the disconnect between the growth numbers and the price decline looks like opportunity, not a trap.