The trading screen went red and stayed red as SpaceX released its first earnings as a public company. You could hear the wince in investors’ accounts when they saw an $18.4 billion spending line. I stayed on the call while Elon Musk promised robots on the Moon.
I’ll be blunt: that $18.4 billion (≈€16.9 billion) capex number rewrites expectations for what SpaceX is now — not just a rocket company, but a cash-burning AI play. You and other investors are asking whether the gamble makes sense, or whether the IPO glow has already faded into a hangover.
The trading screen flashed red the moment the numbers hit the wires.
The headline: SpaceX spent $18.4 billion (≈€16.9 billion) in Q1 and warned of more heavy spending for the next two quarters. That figure sent shares down more than 8% in late trading and left the market staring at a company that is burning cash at a pace usually seen in early-stage startups, not in a freshly public aerospace giant.
What compounds the worry is scale: $15.8 billion (≈€14.5 billion) of that outlay was poured into AI efforts, a division that closed the quarter with a $1.3 billion (≈€1.2 billion) net operating loss. Point-blank: the balance sheet now reads like a mission plan written in a different currency of risk and ambition.
Why did SpaceX stock drop after the earnings call?
Because the market hates uncertainty as much as it loves growth. You had three shocks at once: massive capex, a systemically unprofitable AI unit, and a looming lockup expiration that frees insiders to sell more than $100 billion (≈€92 billion) of shares. Those dynamics compress confidence fast — and when confidence falls, shares follow.
A desk of analysts refreshed price charts until the lockup deadline loomed.
That lockup ending on Thursday matters. When insiders can sell tens of billions, the supply overhang puts immediate downward pressure on price. Add to that the IPO has already retreated from its mid-June highs — the value lost was roughly equal to Tesla’s market cap at one point — and you have a fragile market mood.
I’ll tell you what investors are calculating: what fraction of future returns will come from launches and Starlink, and what will be required to justify plowing most capital into AI instead. That’s the decision frame driving late trading.
The transcript was full of bold claims and timelines.
Musk promised two gigawatts of compute by year-end and “several times higher” cumulative compute by the end of next year. He named Nvidia as the exclusive chip supplier and described a partnership to build Starmind AI satellites that would start launching in 2027.
Directing $15.8 billion toward AI felt like a high-stakes casino bet: enormous potential payoff, but with fierce odds and a short runway for proof. If you’re an investor focused on near-term cash flow, that bet will look irrational; if you’re a futurist chasing scale in compute and edge AI, it might look brilliant.
How much did SpaceX spend on AI in Q1?
SpaceX reported $15.8 billion (≈€14.5 billion) in AI-related capital spending in the quarter, and that arm posted a $1.3 billion (≈€1.2 billion) operating loss. Those two numbers are why the call felt less like investor relations and more like damage control for nervous shareholders.
The call promised satellites, chips, and a chatbot trained on every telemetry stream.
Musk said Grok 5 will arrive before year-end, trained on “all the data that SpaceX has ever produced,” pitched as “by far the best engineer.” He has also described the model as AGI-level on prior occasions — a contested claim in the industry — and plans Starmind satellites as an orbital compute fabric. Nvidia, Grok, LLMs and space-based data centers are now entwined in a single thesis.
That thesis raises immediate questions about cost-per-flop, launch externalities, and regulatory scrutiny. Some experts doubt the economics of orbital data centers; others worry about launch frequency and environmental impact if tens of thousands of payloads become the plan.
When will SpaceX begin launching Starmind AI satellites?
Musk signaled the program would start launches in 2027. If that schedule holds, you’re looking at an unprecedented cadence: building terrestrial compute and then trying to scale it into orbit, with all the logistical, regulatory, and engineering hurdles that implies.
The CEO kept a defiant tone even as analysts pressed for specifics.
Musk pushed back: terrestrial data centers are a “trivial problem” next to reusable rockets, and he doubled down on a daily-launch target next year and robot-staffed factories on the Moon. Those images are cinematic — and that’s both the appeal and the risk. Bold storytelling can rally capital, but it can also inflate expectations that will later be punctured by reality.
The market’s reaction was not a slow burn. Selling pressure can cascade like stacked dominoes once momentum shifts, and the combination of heavy capex, ongoing losses in AI, and a looming insider sell window creates a fragile setup.
So what should you watch next?
Watch the cash runway, the cadence of compute deployment, and any concrete benchmarks for Grok 5’s performance. Track Nvidia supply agreements and whether SpaceX publishes realistic unit economics for orbital compute. Pay attention to insider sales after the lockup; they will tell you more about internal confidence than any prepared slide deck.
I’ve covered speculative capital flows before, and here’s my read: this is not only a technical bet — it’s a credibility test. If the company misses early AI milestones or if the lockup release floods the market, sentiment could swing hard and fast.
Elon has sold grand futures before; now he’s asking the market to underwrite one that layers AI onto rockets. Will investors buy into that layered future, or will they sell and leave the Moon to the rhetoric?