SpaceX Lost $5 Billion Last Year. It Wants a $1.75 Trillion IPO. Here Is Where the Math Gets Uncomfortable.
SpaceX posted a $5 billion loss in 2025 on $18.5 billion in revenue, then confidentially filed for a $1.75 trillion IPO that would make it the largest public offering in history. That is 94.6x trailing revenue, a multiple no major public company has sustained. Inside the subscriber growth, margin expansion, and xAI gamble required to make the numbers work.
Ninety-four point six. That is the price-to-revenue ratio implied by SpaceX's reported $1.75 trillion target IPO valuation against its $18.5 billion in 2025 revenue. For context, Apple trades around 8x. Nvidia, in the middle of the most violent demand cycle in semiconductor history, peaked at roughly 30x. Amazon, during the years when Wall Street gave Jeff Bezos a blank check, rarely exceeded 4x.
SpaceX wants 94.6x. And it lost $5 billion doing it.
On April 9, The Information reported that SpaceX posted a net loss of nearly $5 billion in 2025 on revenue exceeding $18.5 billion. That loss includes Elon Musk's AI startup xAI, which SpaceX acquired in February 2025 in a deal structured as a merger. Just one year earlier, SpaceX had been a profit machine: Reuters reported in January 2026 that the company generated roughly $8 billion in profit on $15 to $16 billion in 2024 revenue.
From $8 billion in profit to $5 billion in losses. A $13 billion swing in 12 months. SpaceX's response: file confidentially with the SEC and ask public markets for the largest valuation ever assigned to a newly listed company.
Revenue Is Real
Before dissecting the valuation, the underlying business deserves credit. SpaceX's revenue trajectory from 2021 to 2024 was one of the most impressive compounding runs in corporate history:
| Year | Revenue | YoY Growth | Net Income |
|---|---|---|---|
| 2021 | $2.3B | - | Not disclosed |
| 2022 | $4.6B | +100% | Not disclosed |
| 2023 | $8.7B | +89% | Not disclosed |
| 2024 | ~$15.5B | +78% | ~$8B profit |
| 2025 | $18.5B+ | ~19% | -$5B loss |
Revenue sources: Motley Fool and Quilty Space. Net income sources: Reuters.
Starlink subscribers doubled from roughly 4.6 million at end of 2024 to over 9.2 million by end of 2025, reaching 10 million by February 2026. Starlink operates in 155+ countries. Maritime revenue grew 55% year-over-year. Aviation added 1,400 aircraft in a single year, up from 450 in 2024. Direct-to-Cell signed 27 mobile network operator partners and connected 12 million users across 22 countries.
Quilty Space forecasts Starlink alone will generate $20 billion in 2026 revenue, $14 billion in EBITDA, and $8.1 billion in free cash flow. Add launch services (Falcon 9 dominated 97% of U.S. orbital mass to orbit in 2025), and the total enterprise revenue likely exceeds $22 billion.
A Multiple Problem
Here is where the admiration ends and the arithmetic begins.
| Metric | SpaceX (at $1.75T) | T-Mobile | Netflix | Nvidia (peak) |
|---|---|---|---|---|
| Price/Revenue | 87.5x forward | ~3.6x | ~8.9x | ~30x |
| EV/EBITDA | ~125x | ~12x | ~35x | ~40x |
| Price/FCF | ~216x | ~20x | ~50x | ~55x |
T-Mobile is the closest functional analog: a wireless carrier with 100+ million subscribers, ~$85 billion in revenue, and a market capitalization around $310 billion. It trades at 3.6x revenue. SpaceX wants 24 times that multiple.
Netflix, the poster child for subscription growth commanding a premium, trades at roughly 9x revenue after a decade of proving it could scale globally while maintaining pricing power. SpaceX wants nearly 10 times Netflix's multiple.
Nvidia, benefiting from an unprecedented GPU demand cycle driven by AI training infrastructure, peaked at roughly 30x forward revenue in mid-2024. SpaceX wants triple Nvidia's peak multiple, while losing money.
Growth Runway Math
Valuations are forward-looking. What matters is not what SpaceX earns today but what it will earn in 5 to 10 years. So let us run the math.
For $1.75 trillion to be a reasonable entry point, SpaceX needs to reach a valuation-justified level of earnings. Mature technology companies typically settle around 20 to 30x EBITDA. Using a generous 25x target multiple, SpaceX needs to generate $70 billion in annual EBITDA to be worth $1.75 trillion on fundamentals alone.
Quilty estimates Starlink's EBITDA margin at roughly 70%. At that margin, $70 billion in EBITDA requires $100 billion in revenue. SpaceX is tracking toward $22 billion in 2026. SpaceX needs to grow revenue by roughly 4.5x from current levels.
How fast can that happen?
| CAGR Scenario | Years to $100B Revenue | Target Year |
|---|---|---|
| 35% (2021-2024 average) | ~5 years | 2031 |
| 25% (optimistic maturation) | ~7 years | 2033 |
| 15% (realistic deceleration) | ~11 years | 2037 |
SpaceX's growth already decelerated from 89% (2023) to 78% (2024) to 19% (2025). Quilty's 2026 forecast implies 8% Starlink revenue growth. If you are buying at $1.75 trillion today, you are betting growth re-accelerates and sustains 25%+ for the better part of a decade. That 19% slowdown in 2025 is not evidence of such a trajectory.
A $39/Month Problem
In March 2026, SpaceX cut its residential entry price to $39 per month (from $50) for the 100 Mbps tier and began offering free equipment rentals, eliminating the $349 upfront cost. SpaceX cited expanding accessibility. Timing, weeks before the IPO filing, suggested subscriber count optimization.
Run the ARPU math. At $39 per month ($468/year) with 10 million current subscribers, that is $4.68 billion in annual consumer revenue from the base tier alone. Quilty's $11.3 billion consumer forecast implies a blended ARPU closer to $900 per year, reflecting higher-tier plans and premium markets. But the $39 promo pushes new subscribers into the lowest-ARPU cohort.
To reach $100 billion in total revenue at current segment mix (roughly 55% consumer, 10% enterprise, 10% maritime, 5% aviation, 20% launch), consumer revenue needs to reach $55 billion. At the current blended $900/year ARPU, that requires 61 million consumer subscribers. At a promo-depressed $600/year ARPU, that requires 92 million.
Starlink currently has 10 million subscribers in a total addressable market of roughly 500 million households globally without reliable broadband. Getting to 60 million is plausible on a decade timeline. Getting there while Amazon's Project Kuiper is deploying its own constellation, with a company that has a track record of subsidizing market entry below cost, is a different proposition.
A $13 Billion Question: What Is xAI Costing?
SpaceX earned $8 billion in 2024. It lost $5 billion in 2025. One structural change explains it: xAI was folded in. That is a $13 billion swing attributable, at least in large part, to absorbing an AI startup in its capital-intensive growth phase.
Whether explicitly stated or not, IPO investors are pricing in xAI. SpaceX is no longer just rockets and satellites. It is rockets, satellites, and a bet on orbital AI infrastructure. SpaceX has discussed deploying AI data centers in orbit, leveraging its own launch capabilities to reduce deployment costs.
This is either visionary or a capital sinkhole. Investors draw the AWS circa 2005 comparison: Amazon was unprofitable in aggregate because it was funding cloud infrastructure that would eventually generate more profit than its retail business. But AWS had paying enterprise customers from day one. xAI's revenue contribution to SpaceX is not publicly known, but the $13 billion swing implies it is consuming capital faster than it is generating it.
What Would Justify $1.75 Trillion
SpaceX is not a telecom company, and valuing it like one misses the structural point. SpaceX controls the launch infrastructure (97% of U.S. orbital mass), the broadband constellation (9,700+ satellites, growing by hundreds per month), the direct-to-cell layer (27 MNO partners, 650 satellites in 18 months), and now AI compute. It manufactures its own rockets, its own satellites, its own user terminals, and is building its own ground stations.
This is vertical integration at a scale that has no historical precedent in aerospace or telecommunications. Amazon traded at over 100x earnings for more than a decade because investors correctly identified it as a platform monopoly whose infrastructure spending would generate compounding returns. SpaceX's Starship, if it reaches routine operation, adds a physical barrier to competition that cannot be replicated by software. Building a rival launch vehicle, deploying a competing 10,000-satellite constellation, and signing 27 mobile operators takes a minimum of 8 to 10 years and tens of billions of dollars. Amazon Kuiper has spent 6 years and billions to reach zero commercial subscribers.
If Starlink V3 satellites (1 Tbps downlink each, versus V2's current capacity) deploy on Starship at the promised 60 Tbps per launch, the capacity economics change fundamentally. At scale, SpaceX could offer competitive broadband to every human on Earth, in the air, and at sea, from a single integrated system. No terrestrial carrier can match that geographic reach.
Limitations
This analysis relies entirely on leaked financials, analyst estimates, and press reports. SpaceX is private and has never filed public financial statements. The $18.5 billion revenue and $5 billion loss figures come from The Information, citing unnamed sources. Quilty Space's forecasts are well-regarded in the satellite industry but are external projections, not company guidance. xAI cost attribution is inferred from the timing of the profit-to-loss swing, not from a disclosed segment breakdown. We do not know SpaceX's debt structure, preferred equity terms, or the specific revenue contribution of xAI. Critically, we do not know Starlink's subscriber churn rate. If churn is 3% monthly (common for value-tier ISPs), the net subscriber addition story changes dramatically. Promotional pricing at $39/month may also depress ARPU trends below Quilty's current forecasts, which were modeled on pre-promotion pricing.
What You Can Do
If you are considering buying SpaceX at IPO, build your own DCF model using Quilty's published forecasts ($20B revenue, $14B EBITDA, $8.1B FCF for 2026) as a starting point. Stress-test at 15%, 25%, and 35% revenue CAGRs over 10 years and apply a terminal multiple of 20 to 30x EBITDA. Compare your calculated fair value against $1.75 trillion. If your model requires sustained 30%+ growth for 8 or more years to break even at the offering price, acknowledge you are making a thesis bet on monopoly economics, not a value investment.
If you are a Starlink subscriber or potential customer, the IPO timing benefits you. SpaceX is cutting prices and giving away hardware to boost pre-IPO metrics. At $39/month with free equipment, this is the best deal Starlink has ever offered. Lock in promotional pricing before the company needs to demonstrate ARPU growth to public-market investors.
If you work in the satellite or telecom industry, the IPO prospectus (when it becomes public) will be the first-ever detailed look at Starlink's real unit economics, churn rates, and segment margins. That document will contain more actionable competitive intelligence about the LEO broadband market than every analyst report published in the past five years combined.
Bottom Line
SpaceX built something extraordinary. A vertically integrated space infrastructure company generating $18.5 billion in revenue, doubling subscribers annually, and operating in 155 countries. That is not in dispute. What is in dispute is whether that entity, carrying a $5 billion loss from its xAI merger, with growth decelerating from 89% to 19% in two years, is worth $1.75 trillion at listing. At 87.5x forward revenue, investors are not paying for what SpaceX is. They are paying for what SpaceX might become if Starship works, if V3 satellites deploy on schedule, if xAI generates returns, and if no competitor closes the gap. It can work, but it requires running 25%+ compounding for 7 years to reach justified fundamentals. History suggests that buying at 94x revenue is less about analysis and more about faith. Sometimes faith pays. Saudi Aramco's 2019 IPO at $1.7 trillion looked similarly stretched and now trades near $1.8 trillion. But Aramco sits on 261 billion barrels of proven reserves. SpaceX sits on a constellation of depreciating satellites, a bet on AI it absorbed 14 months ago, and the assumption that a rocket currently in test flight will become as routine as a 737.