$18.7 billion in revenue. $4.9 billion in losses.
SpaceX's S-1 — filed publicly on May 20, 2026 — is the first time the company has disclosed full audited financials. Consolidated FY2025 revenue grew 33% year-on-year to $18.674 billion, anchored by Starlink at $11.4 billion. Net loss widened to $4.9 billion, driven by $3.0 billion in Starship R&D and the absorption of xAI in February 2026. The headline trade is a profitable, fast-growing connectivity business inside a heavy-spend launch-and-AI parent.
Three years, nearly doubled.
Consolidated Revenue
FY2025 is the first year reported on a post-xAI-merger consolidated basis (merger closed February 2026; xAI added retroactively under purchase accounting).
Revenue by segment (FY2025)
Where the money actually comes from.
The S-1 reports four reportable segments under ASC 280. Revenue, segment operating income, and year-on-year growth are disclosed for Connectivity. Detailed margins for Launch, Starshield, and xAI/Other are partially redacted or aggregated to protect national-security customer pricing — a pattern Wall Street Journal and Bloomberg flagged in their initial coverage.
| Segment | FY2025 Revenue | YoY Growth | Op Income | Op Margin |
|---|---|---|---|---|
| Connectivity (Starlink) | $11.4B | ~+50% | $4.4B | ~38.6% |
| Launch services (Falcon) | ~$4.2B | ~+15% | Not split | Not disclosed |
| Starshield | ~$1.8B | +~80% | Not disclosed | Not disclosed |
| xAI / Other | ~$1.3B | n/a (new) | Loss | Loss |
| Total / consolidated | $18.67B | +33.2% | $(2.6)B | (14.0)% |
Connectivity (Starlink)
Starlink revenue of $11.4 billion represented roughly 61% of consolidated top line and grew ~50% year-on-year. Segment operating income was $4.4 billion, implying an operating margin near 38.6% — well above the 25–30% gross margin range Bloomberg estimated in late 2024 leaks. The S-1 attributes the inflection to per-satellite payload utilization rising as the constellation matured, ground-station depreciation flattening, and direct-to-cell adding minimal incremental network cost. Subscriber adds of 750,000 to 1.5 million per month through Q1 2026 brought the base to 10.3 million across 164 countries.
Launch services
Launch revenue is partially internal — Starlink itself is the single largest consumer of Falcon launches, and intercompany launches are eliminated in consolidation. The external launch line still generated approximately $4.2 billion, anchored by NASA Commercial Crew (CRS-2), NRO national security launches, and a long list of commercial constellation customers. The S-1 notes Falcon held ~90% global commercial launch share in 2025 by mass to orbit.
Starshield
Starshield revenue of roughly $1.8 billion is the second-fastest-growing segment but with the least disclosure. The S-1 confirms multi-year contracts with the National Reconnaissance Office and U.S. Department of Defense for a classified Earth-observation constellation built on Starlink V2 chassis. Margin is described as "consistent with prime defense contractors" — language analysts have read as low-double-digit operating margin.
xAI / Other
The xAI segment captures Grok model API revenue, X platform subscription revenue (X Premium, X Premium+), and early commercial revenue from orbital AI research. The S-1 reports the segment ran at an operating loss in FY2025 — compute capex and Memphis cluster expansion outpaced model API revenue.
Where the $4.9B loss came from.
The headline loss is the product of three forces: very high R&D (Starship plus xAI compute), heavy non-cash stock-based compensation, and capacity capex still ramping ahead of revenue.
| Line item | FY2025 | % of revenue | Comment |
|---|---|---|---|
| Revenue | $18.67B | 100.0% | Three-segment growth plus xAI add |
| Cost of revenue | ~$11.4B | ~61% | Launch operations, satellite depreciation, ground network, cloud compute pass-through |
| Research & development | ~$6.8B | ~36% | Of which Starship R&D = $3.0B; xAI/compute = ~$1.9B; balance Starlink V2 + Raptor |
| Selling, general & admin | ~$1.9B | ~10% | Sales force for Starlink Business + Direct to Cell, plus legal/compliance build-out for IPO |
| Stock-based compensation | Material (non-cash) | — | Disclosed inside the lines above; large at IPO-readiness companies |
| Operating income | ~$(2.6)B | ~(14)% | Excludes interest, taxes, FX |
| Net loss | $(4.9)B | (26)% | Reported figure per S-1 |
The capital intensity nobody else can match.
SpaceX is one of the most capital-intensive non-utility businesses in the public market. Even before the IPO, the company spent multiple billions a year on three concurrent build-outs:
- Starship. The Boca Chica build site, Raptor engine factory, and orbital launch and recovery infrastructure consume most of the headline R&D number. Per the S-1, Starship R&D was $3.0B in FY2025 and is running at a ~$3.7B annualized pace in Q1 2026.
- Starlink V2 satellites. Each V2 generation roughly doubles per-satellite throughput. Production is run vertically at Hawthorne and Bastrop. The S-1 footnotes capitalized satellite costs at "tens of millions" per unit at scale.
- Ground stations & gateways. Ground network expansion is funded out of operating cash flow and adds capacity ahead of subscriber growth — particularly for direct-to-cell.
Use-of-proceeds language in the S-1 directly funds these three lines, plus xAI compute. After the IPO, SpaceX will become — for the first time in its history — funded from public capital markets at scale. Reuters and Bloomberg have reported the company has held more than $4 billion of cash on hand throughout 2025, but the IPO step-up to a balance-sheet position in excess of $80 billion (post-raise + retained cash) materially changes the company's ability to absorb Starship cost overruns.
When does SPCX show GAAP net income?
The S-1 does not provide forward guidance — public-company IPO documents almost never do. The forward-looking statements section is restricted to qualitative language about "achieving profitability as Starlink scale and Starship economics mature." Read literally, the S-1 implies that GAAP profitability is plausible once Starship launches at scale, and Bloomberg and TechCrunch coverage in late May 2026 quoted unnamed bankers placing first GAAP profitability in the 2027–2028 window.
That timeline depends on three things:
- Starship reusability reaching cost parity with the current Falcon 9 second stage. This compresses launch cost-of-revenue.
- Starlink V2 reducing per-subscriber cost to serve as fleet utilization rises. Connectivity is already at 38% operating margin — incremental gains compound.
- xAI moving from a compute-burning research segment to a paid-product segment with stable gross margin. This is the single weakest link in the model.
Investors should be careful: nothing in the S-1 commits the company to a profitability date. Anyone quoting a year is extrapolating, not citing.
Where SPCX would trade relative to the largest names.
At the $1.75 trillion target valuation, SPCX would carry the following key ratios on FY2025 numbers:
| Ratio | SPCX at $1.75T | Method |
|---|---|---|
| Price / Sales (trailing) | ~93.7× | $1,750B ÷ $18.67B revenue |
| Price / Earnings (trailing) | n/a (loss) | Net loss of $4.9B |
| EV / Revenue | ~93× | Approx; before raise applied to cash balance |
| EV / EBITDA | n/a | Negative EBITDA |
| Price / Book | Not yet calculable | Post-IPO book value pending |
Comparable mega-cap multiples
| Company | Approx P/S | Note |
|---|---|---|
| NVIDIA (NVDA) | ~30–35× | Profitable, ~60% net margin |
| Tesla (TSLA) | ~10–12× | Profitable, mature auto cycle |
| Amazon (AMZN) | ~3–4× | Profitable, low-margin retail blend |
| Alphabet (GOOGL) | ~6–7× | Profitable, advertising-led |
| Microsoft (MSFT) | ~12–14× | Profitable, recurring SaaS |
| SpaceX (SPCX) target | ~94× | Unprofitable, growing 33% YoY |