SpaceX's First Earnings as a Public Company — The Numbers Wall Street Wasn't Ready For

On June 12, 2026, SpaceX opened on the Nasdaq at $150 per share in what became the biggest IPO in history. On June 16, the stock peaked at $225.64 — a 50% gain from IPO price in four trading days, driven by the same AI infrastructure euphoria that had been propelling every company in the sector to new heights. Then reality arrived. By August 4, when SpaceX reported its first-ever quarterly earnings as a public company, the stock had retreated to $114 — down almost exactly 50% from its peak, and below the $135 IPO price. The company that broke IPO records was, within eight weeks of going public, trading in the red for everyone who bought at the offering.

Then SpaceX reported its Q2 numbers. The stock jumped 9.4% in a single session — its best day since June 15. Here is what the numbers actually say, what they mean, and why the story of SpaceX as a public company is more complicated and more interesting than either the IPO hype or the post-IPO crash suggests.

The IPO Hangover — What Happened to the Stock

SpaceX's stock trajectory since its June 12 IPO is a case study in what happens when the most anticipated public offering in technology history meets the realities of quarterly reporting. The euphoria was real: SpaceX opened 11% above its $135 IPO price on day one, surged to a record close of $211 within days, and briefly traded above $225 — a level that implied the company was worth more than most of the S&P 500. Then the AI infrastructure investment wave began cooling. Several high-profile AI companies reported earnings that showed costs rising faster than revenue. Investor sentiment toward massive, pre-profitability AI bets shifted from uncritical enthusiasm to measured scrutiny.

SpaceX found itself in the uncomfortable position of a company whose story — rockets, Starlink, orbital data centres, xAI, Terafab, Starship, Grok — is genuinely extraordinary, but whose near-term financials reflect a company burning billions to build the infrastructure for a future that has not arrived yet. The S-1 filing SpaceX submitted for the IPO had already disclosed that the company lost roughly $4.9 billion in 2025 and reported a loss of $4.3 billion in Q1 2026. For investors who had absorbed those disclosures, the post-IPO decline was not a surprise. For retail investors who bought on IPO day hype without reading the filing, it was a painful education.

The Business: Starlink Carries Everything

The clearest thing the earnings report revealed is the business structure that makes SpaceX function. The Connectivity segment — Starlink — generated $11.4 billion in revenue in the reporting period. This is the cash engine. Every rocket launch, every AI data centre investment, every Terafab dollar, and every Grok training run ultimately traces back to the 12 million subscribers paying an average of $66 per month for satellite internet.

That subscriber figure is significant: 12 million Starlink subscribers, doubled from a year earlier and up 17% quarter-over-quarter. The growth rate is healthy. The revenue-per-user trajectory is less encouraging: average revenue per user has dropped from $85 a year ago to $66 today, as SpaceX has pursued volume growth in price-sensitive markets — rural users, developing markets, emergency responders, maritime and aviation customers — at lower price points than its early residential customers paid. Musk addressed this directly on the earnings call, arguing that Starlink has evolved from a "patchy" backup internet option to a service with "incredible uptime and low latency" capable of serving as a primary internet provider for offices and government agencies — a higher-value use case that commands better pricing over time.

The Burn: AI Is SpaceX's Biggest Loss

The AI segment is where the financial picture gets complicated. In Q1 2026, SpaceX's AI division — which encompasses the xAI acquisition, Grok model development, orbital data centre buildout, and now the Cursor coding platform (acquired as part of the Anysphere deal) — generated $818 million in revenue while posting an operating loss of $6.36 billion. The revenue comes primarily from infrastructure rental payments from AI companies — Anthropic, Google, and reportedly others — who are paying SpaceX to use its compute capacity as an alternative to the overcrowded and supply-constrained hyperscaler market.

The $6.36 billion operating loss reflects the cost of building that compute infrastructure at the pace SpaceX is targeting. Total capex was $10.1 billion in Q1 alone, of which approximately $7.7 billion was allocated to AI-related investments. To put that in context: SpaceX spent more on AI infrastructure in a single quarter than many large companies spend in a year. Analysts do not expect the AI segment to turn profitable in the near term — Morningstar's Nicolas Owens said plainly: "I don't expect AI to be profitable for a long time." The question is whether the infrastructure being built today will generate returns at the scale required to justify the investment.

The CFO's Bold Claim — $100B ARR by Year-End

SpaceX CFO Bret Johnsen made the earnings call's most striking statement: the company is "on pace to reach $100 billion in annualised recurring revenue by the end of 2026." This is a number that, if achieved, would make SpaceX's revenue run rate comparable to some of the largest technology companies in the world — companies that have been operating for decades and that SpaceX went public alongside just two months ago.

The path to $100 billion ARR runs almost entirely through Starlink. At 12 million subscribers paying $66 per month, Starlink's annualised run rate is approximately $9.5 billion — which means SpaceX would need to roughly triple either its subscriber count or its revenue per user by December 31 to hit the target. Johnsen's confidence likely reflects internal projections about enterprise and government Starlink contracts — which carry much higher monthly rates than residential plans — and the continued subscriber growth that has been doubling year-over-year. The $100B target is ambitious. It is not obviously impossible, which is an unusual combination for a number that large.

What the Earnings Didn't Answer

The questions Wall Street most wanted answered going into the earnings call remain only partially resolved. Investors wanted to understand the return on the AI investment — when orbital data centres become profitable, what the margins look like, and how Grok's commercial trajectory justifies $7.7 billion in quarterly AI capex. The earnings provided revenue figures but not the longer-term model that would allow investors to price the AI segment on anything other than faith in Musk's vision.

Starship progress was another area of focus. The flagship rocket — designed for Mars missions and, eventually, the kind of heavy-lift capability Terafab will need to supply orbital data centres — had both hurdles and successes during July. Its path to commercial operational status and the revenue it will eventually generate from government and commercial launch contracts is the second largest unanswered question behind the AI business model.

The Bigger Picture — SpaceX as AI Infrastructure

The most important reframe for understanding SpaceX post-IPO is that it is no longer primarily a rocket company. Connectivity (Starlink) generates more revenue than everything else combined. The AI segment burns more capital than everything else combined. And the long-term strategic ambition — orbital data centres, Terafab chips, Grok as a frontier AI model, Cursor as the developer tool of choice, Optimus robots built with Tesla chips manufactured in Texas — is fundamentally a play on AI infrastructure at a scale that no other company has attempted.

Whether that bet pays off depends on variables that no quarterly earnings report can resolve: whether AI inference demand scales to the degree that orbital compute becomes cost-competitive with terrestrial data centres; whether Terafab's chip manufacturing ramp delivers on time and at the yields required for profitability; and whether Grok becomes a commercially significant AI model or remains a showcase for xAI's engineering ambition. The Q2 earnings report answered the most pressing question — is SpaceX growing revenue? Yes, apparently fast enough to beat estimates. The deeper question — does the AI infrastructure bet make financial sense at $10 billion per quarter of capex? — will take years, not quarters, to answer.