SpaceX: The Numbers Don't Support the Crash Narrative (Buy)

What more did the market want from SpaceX after its debut earnings report?

Revenue beat estimates by 13%, up 92% year over year. Earnings loss came in at a fraction of what analysts feared. Adjusted EBITDA — a rough proxy for operating cash generation before non-cash charges — nearly tripled. And yet the stock spent two months collapsing 50% from its mid-June peak of $225.64, trading below the $135 IPO price as if the company had failed on day one.

I've been baffled by the reaction. The fundamentals that justified the IPO pricing haven't broken. If anything, they've accelerated. The selloff looks less like a thesis correction and more like a supply shock — the first lockup expiry on August 6th, when 911.5 million shares worth roughly $123 billion became eligible to trade for the first time. That's the kind of share flood that overwhelms even the strongest earnings report.

Now the stock has clawed back, trading near $153 as of August 7th, well above the $105-110 floor it established in late July. The question for investors isn't whether SpaceX is the company its prospectus promised. The question is whether the market has once again misread the risk/reward.

The Fundamentals: Growth at 92%, Not Zero

Let's start with what actually happened in Q2 2026. SpaceX reported $7.81 billion in revenue against a consensus estimate of $6.93 billion. That's not a marginal beat — it's a meaningful one, and it came on top of 92% year-over-year growth. The net loss was $541 million, or $0.09 per share, compared to analyst expectations of a $0.26 loss. The company expected to lose about $1.9 billion that quarter; it lost less than a third of that.

The real story lives in the segments. Starlink, the satellite connectivity division, brought in $4.29 billion and generated $1.66 billion in operating profit — a 38.6% margin. Subscribers doubled to 12 million. This isn't a speculative burn machine anymore; it's a cash-generating business with a structural advantage that no terrestrial broadband provider can replicate at global scale.

The AI segment, which folded in the xAI acquisition, delivered $2.56 billion in revenue, up 247% year over year. More importantly, adjusted EBITDA turned positive at $1.15 billion, flipping from a $609 million loss the prior quarter. That's the kind of inflection the market loves — until the capex number comes out.

And here's where the crash narrative found its fuel.

The Capex Question: $18.4 Billion in a Quarter

SpaceX spent $18.37 billion on capital expenditures in Q2 alone. That's 235% of quarterly revenue. Eighty-six percent of it went to AI infrastructure — data centers, compute clusters, the physical backbone of the cloud services business. CFO Bret Johnsen indicated that Q3 and Q4 spending should be similar to Q2 levels, which would push full-year 2026 capex toward $65 billion.

I don't want to minimize this. No, I want to put it in the frame the market is missing.

First, the company ended Q2 with approximately $100 billion in cash, cash equivalents, and marketable securities. Total debt sits at $38.4 billion. SpaceX is not leveraged to the gills; it's sitting on a fortress balance sheet that can absorb the build-out.

Second, the AI capex is buying contracted revenue. The company closed $14.1 billion in contracted cloud-services sales during the quarter, including $6.7 billion contracted for early Q3 alone. The Anthropic deal alone pays SpaceX $1.25 billion per month through 2029. This isn't capex aimed at a hypothetical future customer base — it's capex backed by signed, multi-year contracts.

Third, and most important for the risk/reward calculation: the market has treated these capex numbers as evidence that SpaceX will never be profitable. The S&P projection of negative free cash flow through 2029 is the doomsday anchor. But adjusted EBITDA was $3.54 billion in Q2, beating the $2.10 billion estimate by nearly 70%. If AI infrastructure spending eventually plates off and Starlink margins keep expanding, the cash flow inflection doesn't need to wait until 2030. It needs to happen when the spending curve bends, not when revenue stops growing.

The Moat Check: Why the Pressure Isn't Cracking Anything

Before I call a contrarian setup, I need to verify the competitive moat survives the stress. This is where the falling knife becomes a falling opportunity — or stays a knife.

Starlink's moat is infrastructure-based. Twelve million subscribers, a fleet of thousands of low-earth-orbit satellites, and the manufacturing cadence to keep adding them. No competitor has the launch cost advantage to build a comparable constellation at scale — because SpaceX itself owns the launch capability. The vertical integration is the moat. It's not just a satellite company; it's the only company that designs, builds, and launches its own satellites at a fraction of the cost.

The Space segment, despite an operating loss of $542 million driven by Starship R&D, sits on $47.5 billion in backlog. That's contracted business waiting to be executed. The government contracts alone, including over $6 billion in new multi-year Starshield deals, provide revenue visibility that most defense contractors would kill for.

On the AI side, the moat is less proven and more capital-intensive. But the $14.1 billion in contracted cloud sales and the Colossus compute center already leased to Anthropic give xAI an entry point that most AI infrastructure plays don't have — guaranteed demand from day one. The operating loss of $1.26 billion is real, but the sequential improvement from a $2.47 billion loss in Q1 tells you the trajectory, not a dead end.

None of these moats is cracked. They're under construction, but the market priced them as broken.

The Valuation Disconnect

At $153 per share, SpaceX trades at roughly $1.9 trillion in equity value. Annualized against Q2 revenue of $7.81 billion, that's about 24.4 times trailing twelve-month sales. Against 2025 full-year revenue of $18.7 billion, it's closer to 10x.

If revenue continues compounding at the 92% pace Q2 demonstrated, forward sales for fiscal 2027 would approach $37 billion, pushing the forward P/S below 5x. That's not expensive for a company with a Starlink division generating 38% operating margins and an AI division turning positive on an EBITDA basis.

Compare this to the IPO pricing. The company went public at $1.75 trillion — a valuation the market enthusiastically accepted on the strength of a prospectus, not a single public quarter of results. Now that the results are in, and they're materially stronger than expected, the market has arguably baked in a doomsday narrative built entirely on capex magnitude and share supply concerns.

The valuation disconnect is the opportunity.

The Price Action: Bear Trap or Still Falling?

The chart tells a story that deserves attention. SpaceX formed a double-bottom near $105-110 in late July, bounced hard off it, and has rallied through the $135 IPO level to test the $150 area. The $105-110 zone acted as a floor after the lockup expiry, despite the $123 billion in newly tradable shares. Selling pressure that doesn't break a double-bottom after a 50% decline is the textbook signature of exhaustion.

I don't think investors need to chase the current move. The $153 level is approaching the zone where the risk/reward gets less asymmetric. But the setup remains constructive for accumulation on any pullback toward the $118-120 support area — or even back toward $105 if market-wide sentiment drags it down.

The first lockup tranche has passed. The founder and core investor stake of 46.1% remains locked until June 2027. The rolling 7% tranches over the coming months are material but manageable against a float that's already absorbed one of the largest supply shocks in market history without breaking below $105.

The Dalio Warning — and Why It Doesn't Change the SpaceX Thesis

Ray Dalio, founder of Bridgewater Associates, has been vocal about the AI bubble. On August 4th — the same day SpaceX reported earnings — Dalio told the audience of The Diary of a CEO that we're in classic bubble territory, drawing parallels to 1929 and 2000. He flagged SpaceX specifically as part of the speculative issuance surge, pointing to projected negative free cash flow and governance risks around Musk's 85% voting control.

Dalio's macro bubble thesis is worth hearing. It's not worth applying blindly to a specific company whose numbers he hasn't digested.

SpaceX isn't a theoretical AI startup pricing itself off a TAM estimate. It's a company with $7.81 billion in quarterly revenue, 92% growth, a $100 billion cash balance, $47.5 billion in backlog, and a connectivity business that's already generating $1.66 billion in quarterly operating profit. The AI business, for all its capex intensity, is backed by signed contracts worth more than the entire company's annual revenue a year ago.

Bubbles burst when paper wealth gets converted to real losses. SpaceX's wealth isn't paper — it's rockets that fly, satellites that generate revenue, and compute centers that have paying tenants. The governance risk around Musk's concentrated control is real and worth monitoring, but it's a board structure issue, not a thesis-breaker.

The Action

I'm calling this a Buy, but with discipline.

The fundamentals support a higher valuation than the market has assigned over the past two months. The capex burn is aggressive but funded by a fortress balance sheet and contracted demand. The moat — particularly Starlink's vertical integration — is durable, not cracking.

The better risk/reward sits on weakness. A pullback toward $118-120 offers a more asymmetric entry than chasing the $153 level. If the stock breaks back below $105, I'd reassess whether the double-bottom has failed and wait for a new base to form. But as long as it holds above $118, the contrarian thesis stands.

I would also reassess if Starlink's operating margin deteriorates below 30%, signaling pricing pressure or cost escalation I haven't seen in the numbers yet. Or if the AI segment's contracted revenue pipeline stalls — that's the one moat that needs validation, not assumption.

The market threw SpaceX out with the lockup bathwater. The earnings report proved it deserved better. Don't let this buying opportunity go to waste on a deeper pullback.