SpaceX Nearly Doubles Revenue in First Public Earnings Report — But Shares Fell. Here's Why.
SpaceX reported near-doubling year-over-year revenue in its first quarterly release as a public company, yet shares dropped in after-hours trading as investors zeroed in on the massive spending tied to the company's AI infrastructure build-out.
The Q2 2026 results, covered in a live updates stream by CNBC, marked the first financial disclosure since SpaceX's June 2026 IPO. The offering priced at a maximum of $162.00 per share in a prospectus approved by Bafin on June 5, 2026 (SpaceX EU Prospectus). It was the largest initial public offering on record (New York Times).
The revenue acceleration is the headline. But the sell-off tells you what the market actually heard: SpaceX plans to build its AI data centers exclusively on Nvidia chips and expects to exit 2026 with over 2 gigawatts of compute capacity, according to the same CNBC coverage. Two gigawatts is roughly the output of two large nuclear reactors, all dedicated to running AI models. Those numbers carry direct implications for free cash flow — the cash a company has left after paying for its buildings, equipment, and operations — and for margin compression, which is when rising costs eat into profitability well beyond any single quarter's revenue beat.
The AI strategy did not emerge from nowhere. In February 2026, SpaceX announced it had acquired xAI, folding Musk's AI business into the rocket and satellite operator (Associated Press). Prior to that acquisition, xAI's Grok chatbot was hosted on Microsoft's cloud platform (Associated Press). By bringing AI in-house, SpaceX vertically integrated the compute layer it previously rented, aligning with its own claim of being the only vertically integrated company capable of a scalable AI solution (SpaceX).
That acquisition was not universally welcomed. Some SpaceX investors protested the xAI deal, calling it a bailout and unethical (Associated Press). The tension is worth noting: the same deal that positioned SpaceX to capture AI compute revenue also saddled the pre-IPO company with significant integration risk and governance questions just months before the public offering.
The IPO filing itself revealed the depth of current losses, disclosing a $4.28 billion quarterly loss (Reuters). Against that backdrop, the revenue doubling in Q2 is directionally positive but insufficient to offset investor concern about the forward spend trajectory, which is why the after-hours tape traded down despite the strong top-line print.
Musk laid out the demand-side case on his X account on August 4, 2026: Reflection AI will pay SpaceX roughly $150 million per month, approximately $1.8 billion per year, in contracted compute revenue from July 2026 through 2029, contributing to total annualized contracted compute revenue of roughly $28 billion (Elon Musk on X). If those contracted figures hold, they begin to contextualize the capital expenditure. The gap between near-term losses and contracted forward revenue is where the valuation debate will center.
On the infrastructure side, site preparations for SpaceX's Gigabay facility in Florida are underway, with construction targeted for completion and operational status by the end of 2026 (SpaceX). Further out, the company's Starmind AI satellite facility is planned to enable rapid production and deployment of thousands of AI satellites starting as soon as late 2027 (SpaceX). Musk has stated the IPO will bankroll the launch of up to 1 million data-center satellites designed to bypass power and water constraints on Earth (Reuters). That orbital data-center concept faces the same category of hurdles that led Microsoft to abandon a prior terrestrial orbital project, as the same Reuters report noted.
The broader context here is that the market is digesting two competing narratives simultaneously. The first is a company whose quarterly revenue nearly doubled year over year with a contracted compute revenue pipeline in the tens of billions. The second is a company burning over $4 billion per quarter at the pre-IPO filing stage, planning to deploy over 2 gigawatts of Nvidia-exclusive compute by year-end, building satellite factories that will not produce until late 2027 at the earliest, and pursuing an orbital data-center vision with unresolved technical and regulatory feasibility questions. The after-hours decline reflects the market pricing the second narrative as the dominant near-term driver.
For investors and analysts tracking the SPcX tape, the next data points that matter are concrete: realized compute revenue in Q3 and Q4 against Musk's $150 million monthly Reflection AI figure, capital expenditure guidance relative to the 2-gigawatt target, and any progress indicators on Gigabay operational status before year-end. Until the contracted compute revenue shows up in reported financials rather than social media announcements, the gap between the bull case and the income statement remains the core tension.
