In short
SpaceX’s AI revenue more than tripled to $2.6 billion as it sold compute to other AI companies, including Anthropic and Google. But heavy spending on infrastructure and Starship kept the company in the red overall.
- AI revenue climbed to $2.6 billion, more than tripling year over year.
- New compute deals with Anthropic and Google helped drive the increase.
- SpaceX still lost $143 million overall, while its AI division lost $1.5 billion.
- Capital expenditures rose to $18.37 billion as Starship and other projects demanded more spending.
- Starship remains central to Starlink’s future expansion and SpaceX’s long-term strategy.
SpaceX’s artificial intelligence revenue jumped to $2.6 billion in the latest quarter, more than tripling from a year earlier, but the company still posted an overall loss and its AI arm remained deeply unprofitable. The growth was driven largely by new cloud-compute deals with major AI firms, including Anthropic and Google, as SpaceX increasingly behaves like a neocloud provider alongside its core space business.
The results show how Elon Musk’s company is expanding beyond rockets and satellites into one of the most lucrative parts of the AI boom: selling computing power. They also underline the scale of spending required to keep that strategy moving, with capital expenditures climbing sharply and SpaceX’s Starship program continuing to absorb significant resources.
AI revenue leaps as SpaceX leans into compute
SpaceX disclosed that its AI-related revenue rose to $2.6 billion, a gain of more than 200% year over year, according to quarterly earnings materials. Much of that increase came from deals to supply computing capacity to other AI companies, not from consumer products or software in the usual sense.
That puts SpaceX in a fast-growing category of infrastructure providers often described as neoclouds. These companies sell access to computing power, especially GPU-heavy capacity, to AI developers that need enormous resources to train and run models. The business model is attractive because demand is strong, but it also requires major up-front investment in data centers, networking and energy.
For SpaceX, the move represents an unusual diversification. The company is known primarily for rockets, launch services and Starlink, but its AI division has become important enough that, in filings tied to its public-market ambitions, management described it as the source of most of the company’s value.
How SpaceX is competing in the neocloud market
SpaceX is competing by packaging its infrastructure for external AI customers, effectively turning part of its compute capacity into a revenue-generating utility. In May, the company struck a deal with Anthropic. In June, it added another with Google. Those agreements place it in closer competition with other infrastructure players such as CoreWeave, which has built its reputation on serving AI workloads.
The broader significance is that SpaceX is no longer just a buyer of advanced hardware and engineering talent; it is becoming a seller of computational horsepower. That matters because the AI infrastructure market has become one of the hottest segments in tech, with hyperscale demand giving well-capitalized providers a chance to monetize scarce compute capacity at premium rates.
SpaceX’s filings suggest the company sees AI infrastructure as a central part of its future value, even though the business is not yet profitable and is adding to the company’s overall spending burden.
Why is SpaceX still losing money?
SpaceX is still not in the black because the costs of building out its technology stack remain enormous. The company reported a quarterly loss of $143 million, a narrower deficit than the year before, but the result still shows that revenue growth has not yet caught up with the spending needed to support both AI expansion and space development.
The AI division itself lost $1.5 billion during the quarter. That was slightly better than the same period a year earlier, yet it remains a heavy drag on the company’s finances. The combination of high capital spending and operating losses suggests SpaceX is making a deliberate bet that near-term profitability can be sacrificed for long-term dominance in multiple strategic industries.
One key reason for the red ink is that the company is funding more than one capital-intensive frontier at the same time. Building AI infrastructure is expensive on its own. Building launch systems and satellite constellations is expensive as well. Doing both at once magnifies the pressure on cash flow and margins.
| Metric | Latest quarter | Year-ago comparison | What it means |
|---|---|---|---|
| AI revenue | $2.6 billion | More than 3x higher | Rapid growth driven by compute deals |
| AI division loss | $1.5 billion | Slightly smaller | Still a major cost center |
| Total company loss | $143 million | Wider a year earlier | Overall loss narrowed |
| Capital expenditures | $18.37 billion | Higher than prior year | Heavy investment in infrastructure and spacecraft |
| Starship spending | Up $389 million | Compared with last year | Space program remains a major drain |
What is driving the surge in spending?
The biggest drivers are capital expenditures and Starship development. SpaceX said capex reached $18.37 billion, a level that signals a company still in intense build-out mode. These investments are necessary to support both the AI infrastructure push and the company’s larger space ambitions.
Starship is especially important because it is not just another experimental rocket program. It is central to Musk’s broader vision for Starlink and future launch capability. The heavy-lift vehicle is needed to carry more capable, larger satellites that could expand the economics of SpaceX’s internet business, which remains the company’s only profitable segment.
The company said some of the satellites intended for this next phase have already been built, and 20 of them have been launched. Even so, the timeline for complete deployment remains uncertain, and SpaceX did not provide a clear schedule for rolling out 60 satellites at once, which would be an important step toward the full system it wants to field.
How Starship affects the Starlink business
Starship affects Starlink because the rocket has to be capable of lifting heavier satellite versions that can improve the network’s performance and economics. In theory, more capable satellites could strengthen Starlink’s competitive position, increase capacity and support broader global connectivity ambitions.
In practice, that means continued spending before those benefits arrive. Starlink may already be SpaceX’s strongest cash engine, but its future growth depends on a launch system that is still being developed, tested and refined. That makes Starship both a strategic necessity and a financial risk.
What do the earnings say about Musk’s larger AI ambitions?
The earnings offer a clearer picture of Elon Musk’s strategy: SpaceX is not merely funding space exploration while AI happens elsewhere in his portfolio. Instead, the company is shaping itself around the same compute-intensive logic that has defined the AI industry’s winners and near-winners.
That includes the kind of sweeping ambition Musk is known for. His public vision has ranged from data centers in space to a future addressable market far larger than the U.S. economy. Whether those goals are realistic is another matter, but the latest numbers show that parts of that strategy are already translating into revenue.
Still, the figures also show the limits of hype. Even with AI revenue surging, SpaceX remains loss-making overall. The company is earning more from the AI boom, but it is also spending more to participate in it, and the economics are not yet close to balancing out.
Why are AI companies buying capacity from SpaceX?
AI firms are buying capacity from SpaceX because demand for compute remains intense and supply is constrained. Training and serving large models require expensive hardware, reliable power and large-scale data-center infrastructure. Providers that can deliver those resources quickly are in a strong position to sign lucrative contracts.
Anthropic and Google are the clearest examples of that demand in action. By tapping SpaceX for compute, they gain another source of capacity in a market where access to high-end infrastructure can become a competitive advantage. For SpaceX, the contracts represent a way to monetize assets beyond launch services.
This is also part of a broader industry shift. The AI race is no longer only about model quality. It is also about who controls the chips, facilities and network links that make those models possible. Companies that can offer reliable infrastructure are increasingly as important as the companies building the models themselves.
SpaceX’s place in the AI supply chain
SpaceX is not yet a household name in AI infrastructure in the way Nvidia or Microsoft might be, but its earnings suggest it is trying to become a meaningful participant. Rather than selling software or model access, it appears to be supplying the underlying compute that makes those tools possible.
That could be strategically useful because infrastructure customers are often locked in by performance, pricing and capacity availability. But it also means SpaceX has to keep investing aggressively just to stay relevant in a market where faster chips and larger facilities quickly raise the bar.
How does SpaceX compare with other neocloud providers?
SpaceX’s approach resembles other neocloud companies, but it differs in one crucial way: the compute business is being built alongside a massive aerospace operation. That makes the company less focused than dedicated cloud providers, yet it also gives it a broader base of assets and ambitions.
CoreWeave is the most obvious comparator in the source material. Like CoreWeave, SpaceX is selling compute to AI firms that need scalable infrastructure. But unlike a pure-play provider, SpaceX must also fund rockets, satellites, launch systems and a constellation business with global ambitions. That diversification may help over the long run, but it also makes the balance sheet harder to manage.
The table below shows the key elements shaping the comparison.
| Company type | Main offering | Revenue driver | Core risk |
|---|---|---|---|
| SpaceX | Launch, satellites, compute | AI compute contracts and Starlink | Very high capital intensity |
| CoreWeave | AI cloud infrastructure | GPU rentals and cloud services | Dependency on AI demand and hardware supply |
| Anthropic | Foundation models | Model subscriptions and enterprise use | Compute costs and competition |
| Cloud and AI services | Enterprise cloud and platform services | Scaling infrastructure profitably |
What happened to SpaceX shares?
SpaceX’s shares fell after the report, despite an initial jump that reflected investor enthusiasm over the better-than-expected results. According to Bloomberg, the company beat analyst estimates, but that was not enough to sustain the market’s early optimism.
The reaction suggests investors are still focused on the same central tension that runs through the quarterly report: growth is strong, but costs are stronger. A business can post eye-catching revenue numbers and still fail to convince the market if the path to profitability remains unclear.
That dynamic is especially pronounced for SpaceX because investors are not just betting on one business line. They are effectively underwriting a portfolio of long-duration projects, from AI infrastructure to launch vehicles to satellite internet. Each may have a large future payoff, but each also requires steady capital today.
Timeline of SpaceX’s latest AI push
The pace of SpaceX’s recent AI expansion has been quick, with major customer wins arriving within weeks of one another. The sequence below shows how the company moved from a broader compute strategy to specific commercial partnerships.
| Date | Event | Why it matters |
|---|---|---|
| May | Deal announced with Anthropic | Marked SpaceX’s entry into direct AI compute supply |
| June | Deal announced with Google | Expanded the company’s AI infrastructure customer base |
| Latest quarter | AI revenue reaches $2.6 billion | Shows explosive near-term growth |
| Latest quarter | Capital expenditures hit $18.37 billion | Signals massive investment requirements |
| Latest quarter | 20 next-generation satellites launched | Shows Starship-related progress, but not full deployment |
What should investors watch next?
Investors should watch whether SpaceX can keep growing AI revenue without letting costs overwhelm the rest of the business. The key questions are whether compute contracts expand, whether margins improve, and whether Starship continues to progress toward the satellite-lifting capability needed for Starlink’s next phase.
Another important indicator will be whether the company can turn its AI infrastructure into a durable business line rather than a temporary boost. If SpaceX can convert demand from major AI firms into repeatable, high-margin revenue, its neocloud strategy could become a serious asset. If not, it may remain an expensive side bet inside an already costly empire.
For now, the latest quarter points to both promise and strain. SpaceX is making meaningful money from the AI boom, but it is also spending heavily to stay in the game. That combination may satisfy long-term believers, but it leaves the company far from the kind of earnings profile public-market investors usually prefer.
The bigger picture
SpaceX’s results reflect a broader shift in tech: the AI economy increasingly rewards those who control physical infrastructure, not just algorithms. Chips, power, facilities and launch systems are becoming strategic assets in their own right. SpaceX is trying to own a piece of that stack while continuing to pursue the far more capital-intensive dream of interplanetary transport.
That ambition makes the company one of the most unusual businesses in the market. It is part rocket company, part broadband provider, and now part AI infrastructure vendor. The latest quarter shows that each of those roles can generate real revenue, but also that the cost of building them simultaneously is enormous.
For Elon Musk, the numbers are a reminder that scale alone is not the same as profitability. SpaceX may be winning new business in AI and making progress on its space platform, but it is still paying a steep price for that expansion. The company’s future value may depend on whether those bets eventually start reinforcing one another instead of simply adding up to bigger bills.
Frequently asked questions
How much AI revenue did SpaceX report this quarter?
SpaceX reported $2.6 billion in AI revenue for the quarter. That figure was more than three times higher than a year earlier and was driven largely by contracts to supply compute capacity to other AI companies rather than by consumer AI products.
Why is SpaceX getting into AI infrastructure?
SpaceX is getting into AI infrastructure because demand for computing power is soaring and companies need access to large-scale capacity. By selling compute to firms such as Anthropic and Google, SpaceX can monetize infrastructure in the same way a neocloud provider would.
Is SpaceX profitable now?
No. SpaceX still reported an overall quarterly loss of $143 million. Its AI unit also lost $1.5 billion, and the company’s spending on Starship, satellites and compute infrastructure remains very high despite rapid revenue growth.
What role does Starship play in SpaceX’s business?
Starship is central to SpaceX’s future because it needs a heavy-lift vehicle to launch larger satellites for Starlink. The rocket program also consumes large amounts of capital, making it one of the biggest reasons the company is still spending heavily.
Which AI companies are working with SpaceX?
SpaceX has announced deals with Anthropic and Google to provide compute capacity. Those agreements show the company is competing in the neocloud market and trying to become a meaningful supplier of AI infrastructure.









