SpaceX promised a trillion dollars in revenue by 2030. I don't believe it, and I'll explain why
Elon Musk said it on yesterday's earnings call with complete calm, as if he were talking about the weather: $SPCX's internal projections for reaching a trillion dollars in annual revenue have moved from 2031 to 2030. And then he added the sentence that filled financial headlines around the world—that there is a "non-trivial chance" for 2029.
I'll tell you straight up where I stand. I see it as a marketing number, not a business plan. And the reasons I think so are in the data that SpaceX itself has published.
Quick analysis of the results
SpaceX published its first-ever earnings report as a publicly traded company, and the numbers were strong across all segments.
Total revenue for the second quarter of 2026 reached $7.8 billion, a year-over-year increase of 92%. Analysts had expected around $6.8 to $6.9 billion, so the company significantly beat estimates. The net loss narrowed to $541 million from a billion last year, and adjusted EBITDA nearly tripled to $3.5 billion.
The Space segment earned $962 million, up 29% year-over-year, but remains loss-making. The operating loss was $542 million due to huge investments in Starship development. The company launched 38 missions during the quarter and successfully completed the Starship Flight 13 in July, which for the first time demonstrated a Raptor engine restart in space and the softest landing on the water surface ever recorded.
Connectivity, i.e., Starlink, is the company's clear engine—$4.3 billion in revenue (+66% year-over-year), operating profit of $1.7 billion (+79%), and the number of subscribers doubled year-over-year to 12 million. ARPU remained stable at $66 per month. The segment is mainly driven by enterprise and government business, which jumped 108% year-over-year—new clients include $AAL and $LUV, as well as government contracts for Starshield worth over $6 billion.
The AI segment, which includes Grok and xAI infrastructure, recorded revenue of $2.6 billion, up 247% year-over-year. The operating loss narrowed by 49% compared to the first quarter to $1.3 billion, and the segment achieved positive adjusted EBITDA ($1.1 billion) for the first time. Computing capacity rose to 1.4 GW from 0.4 GW last year.
The most striking figure is capex. In a single quarter, SpaceX spent $18.4 billion, of which $15.8 billion went exactly to AI infrastructure. That is a pace of spending that would cause panic at most companies. SpaceX can afford it—after the IPO (net proceeds of $85.7 billion) and the issuance of $25 billion in bonds, it sits on over $100 billion in cash and securities. But it shows what the real game is: a bet that AI computing capacity will be so valuable that it pays off to build it faster than anyone else can keep up.
The market, however, reacted to these numbers with a share price decline of more than 7% in after-hours trading—despite the company beating estimates across all segments.
Under what assumption does a trillion dollars even have a chance
Musk's math rests on one critical assumption: that Starlink's monetization can be massively expanded through the new generation of V3 satellites. He said on the call literally: even if the yield per unit of data were to drop tenfold, that would be enough for a tenfold increase in Starlink's revenues, because the capacity of V3 satellites is an order of magnitude higher than that of the current generation.
In other words, the bet is not on Starlink gaining ten times as many customers. It is on the infrastructure itself being able to transmit orders of magnitude more data, and even at a lower price per unit, that will generate much more money. Added to this is the AI segment, where SpaceX has already signed cloud contracts worth $14.1 billion this year and is preparing the acquisition of Cursor for $60 billion to strengthen its position in AI infrastructure and enterprise tools.
So the puzzle looks like this: Starlink as the main source of revenue from data monetization, AI compute as a fast-growing second pillar, and Starship as a long-term bet on reducing the cost of transport to orbit by 99%, which would open entirely new markets, from space-based data centers to resource extraction.
The company also expects to reach an annual revenue run rate of over $100 billion by December of this year. That alone would be an impressive number. But between $100 billion ARR and a trillion in actual annual revenue yawns a chasm that needs to be seen in perspective.
Why I don't believe it
SpaceX had total revenue of $18.7 billion for the full year 2025. Getting from there to a trillion dollars by 2030 means roughly a 54-fold increase in five years, a compound annual growth rate of around 120%. For comparison, even the most aggressive technology companies in history grew at a long-term rate of 40–60% per year, and that only for a limited period before growth naturally slowed down.
Wall Street analysts have much more sober numbers, and I side with them. Goldman Sachs, which led the underwriting of SpaceX's IPO, estimates 2030 revenue at around $470–474 billion, i.e., less than half of Musk's target. Morgan Stanley is even more cautious, at around $330 billion. And for 2029, when Musk talks about a "non-trivial chance" of a trillion, the FactSet analyst consensus gives a mere $207 billion.
The difference between $207 billion and a trillion is not a rounding error. It is a gap that to me means only one thing: Musk is talking about a scenario from the far upper edge of the probability curve, not a realistic estimate.
Musk's numbers are a scenario, not a forecast
When Musk says "non-trivial chance," it does not mean a prediction. It means that his model has a wide enough probability distribution for 2029 to fall into it as an outlier. That is a fundamental difference from how the media headlines it, which is why I approach it with a large grain of salt.
We have known this pattern from Musk for years. Tesla was supposed to have self-driving cars "next year" since 2016. The Cybertruck was supposed to come out in two years; it came out in five. It's not a lie in the classical sense. It's just the way Musk motivates teams and investors to extreme speed. The problem arises when a marketing ambition becomes the basis for valuing the company, and that is exactly what I think is happening with SpaceX right now.
And here it's doubly sensitive. After the IPO, SpaceX is a company whose valuation rests largely on future growth, not on current profits. The stock has fallen more than 50% from its peak, and the market reacted to the results with a drop of over 7% despite the quarterly numbers being objectively excellent. That confirms to me that investors are already far more concerned about the pace of spending ($18.4 billion per quarter) and the question mark over the AI segment than about stories of a trillion in five years.
What seems realistic to me: Starlink is indeed growing at an enormous pace and has a clear path to monetization through enterprise and government contracts. What seems like pure speculation: that a tenfold increase in revenue from the same data unit can be achieved just through better satellites, in an environment where competition (Amazon's Kuiper, Chinese projects) will push prices down, not up. Musk needs too many things to happen all at once and exactly according to plan, and historically he hasn't succeeded in doing that.