Space Exploration Technologies’ upcoming IPO filing and recent customer announcements have sharpened investor focus on its artificial intelligence strategy, positioning the company as more than a rocket operator. In its S-1, the company estimated a total addressable market of $28.5 trillion, with $26.5 trillion tied to artificial intelligence.
Separately, deal updates tied to AI compute capacity are designed to convert that thesis into near-term revenue. The company also faces a central bottleneck for AI expansion—power availability for data centers—which management is trying to address through an orbital infrastructure plan.
Key takeaways
- Strategic shift: Space Exploration Technologies framed the majority of its market opportunity around artificial intelligence, not space hardware alone.
- AI compute deals: It disclosed arrangements that would rent out compute capacity to Anthropic and Alphabet on a monthly basis, supporting an infrastructure-to-revenue model.
- Catalyst: The announcements reinforce market expectations that AI infrastructure is becoming a product category, not just a technology roadmap.
- Implication for investors: The company’s ability to scale space-based data centers could shape long-term growth, but execution and capital intensity remain key risks.
What Space Exploration Technologies signaled in its IPO filing
In its S-1 filing ahead of its initial public offering, Space Exploration Technologies laid out an unusually broad market opportunity. Out of a forecast $28.5 trillion total addressable market, it estimated $26.5 trillion would be associated with artificial intelligence.
The company is still widely categorized by many investors as a space company. However, the filing’s emphasis on AI suggests management wants the market to value its infrastructure and compute capabilities more directly, with space operations acting as a pathway to AI services.
Deals aimed at turning AI infrastructure into revenue
Ahead of the IPO, Space Exploration Technologies became the counterpart in two separate announcements related to AI compute. According to reporting around the announcements, Anthropic agreed to rent out all compute capacity at Space Exploration Technologies’ Colossus 1 data center for $1.2 billion per month.
In a second deal, Alphabet was reported to pay Space Exploration Technologies $920 million per month to rent compute capacity. Together, the agreements are positioned as proof points that Space Exploration Technologies can monetize AI infrastructure on the ground while building towards a larger model that could extend to space-based data centers.
For investors, the key takeaway is not just the size of the contracts, but what they imply: demand for AI compute capacity is translating into long-term outsourcing relationships with operators capable of delivering infrastructure reliably.
Why power constraints could make “space-based data centers” a differentiator
A core issue for expanding AI workloads is access to power. The article noted that, with traditional ground-based data centers, obstacles include water use, noise impacts, and constraints on power grids—factors that can slow or limit new capacity.
According to research cited from The Motley Fool, U.S. data centers consumed an estimated 177 to 192 terawatt-hours of electricity in 2024, or roughly 4% to 5% of all U.S. electricity. The research also pointed to scenario-based projections from the Electric Power Research Institute suggesting consumption could rise to 9% to 17% by 2030, with the updated range described as about 60% higher than the organization’s earlier projections. The cited commentary emphasized that most of that energy usage comes from conventional data centers.
Space Exploration Technologies’ stated workaround is to pursue orbital data centers that can be powered by the sun. The company applied in January to launch one million satellites into orbit that could function as data centers. It also described a concept known as AI1, featuring solar panels spanning 230 feet, with expectations that orbital data centers could be deployed as early as 2028.
This plan is designed to address the limiting factor of power availability that increasingly determines whether additional AI compute can be brought online quickly and at scale.
Upside case and key risks flagged by investors
The company is attempting to establish the infrastructure needed to become a significant participant in AI services. While its AI division accounted for $3.2 billion in revenue in 2025, Goldman Sachs’ forecast—as referenced in the article—projected that AI revenue could rise to $322 billion by 2030, with total revenue forecast at $474 billion.
However, the same materials also highlight volatility and execution risk. Space Exploration Technologies reported a loss of $4.6 billion in 2023, net income of $791 million in 2024, and net losses of $4.9 billion in 2025, according to the figures cited in the article. The company’s AI ambitions are also described as accounting for a substantial share of capital expenditures—$12.7 billion in 2025 for AI spending compared with $3.8 billion for its space segment and $4.1 billion for its connectivity division.
For market participants, that combination of aggressive scaling potential and heavy investment requirements suggests that the stock’s trading path could remain sensitive to milestones, financing needs, and the pace at which space-based infrastructure moves from concept to commercial operations.
What to watch next
Investors are likely to focus on progress toward commercializing orbital data centers and on any updates that extend the compute-capacity partnerships. The company’s IPO timeline, along with broader signals on data center power constraints and AI demand, will be key as markets look for confirmation that near-term revenue traction can translate into long-term infrastructure scale.
