When Elon Musk first began painting a vision of artificial intelligence empowered by the vastness of space, it captured imaginations across the technology sector. The idea of orbital compute, of data centres floating beyond Earth’s atmosphere, seemed like the natural next step for a company that has made a habit of turning science fiction into engineering reality. Yet, as Wall Street analysts and industry observers take a closer look at SpaceX’s current trajectory, a more grounded picture emerges. The company’s most immediate and substantial value, at least for the foreseeable future, lies not in the stars but in the sprawling terrestrial infrastructure it is building to power the artificial intelligence revolution happening right here on Earth.
This distinction between near-term profitability and long-term vision is becoming increasingly clear as SpaceX continues to monetise its compute capabilities through strategic agreements with enterprise clients. The company has already secured three significant compute contracts with artificial intelligence firms, including a notable partnership with Anthropic for its Colossus supercomputer clusters. These agreements are expected to generate approximately twenty-eight billion dollars in annual revenue, a figure that dramatically overshadows earnings from other business segments in the current fiscal year. When placed alongside SpaceX’s capital expenditure of roughly twelve point seven billion dollars on AI infrastructure in 2025 alone, the scale of terrestrial investment becomes unmistakable.
What makes this shift particularly noteworthy is how it reflects broader trends in the artificial intelligence industry. Infrastructure providers, especially those operating large-scale data centres, have emerged as some of the most significant beneficiaries of the AI boom. As businesses and consumers rapidly incorporate AI into everything from software development and robotics to routine activities like shopping and travel planning, the demand for computational power has grown exponentially. SpaceX, with its substantial resources and engineering expertise, has positioned itself at the heart of this expanding market.
The Colossus facilities, which currently provide roughly one gigawatt of compute capacity, rank among the world’s largest artificial intelligence clusters. This scale is not merely impressive in isolation but represents a fundamental shift in how computational resources are being deployed across industries. J.P. Morgan analysts have projected that SpaceX will expand its terrestrial AI compute capacity to approximately nine gigawatts by 2029. To put that number in perspective, it equals roughly four times the power generated by the Hoover Dam, an indication of just how substantial the company’s terrestrial ambitions have become.
Anthony Milovantsev, a partner at the consultancy firm Altman Solon, offers a measured perspective on the company’s dual-track strategy. The narrative that orbital compute will fundamentally disrupt terrestrial data centres is a little bit overblown, Milovantsev observed. Any kind of displacement of terrestrial data centres is ten years plus out. His assessment reflects a growing consensus among industry analysts that while the long-term potential of space-based computing is genuine, the practical challenges of achieving it remain significant.
These challenges are not insignificant. SpaceX’s orbital AI ambitions depend heavily on the successful introduction of Starship rockets, which are still undergoing development and testing. Lower launch costs, which are essential for making orbital compute economically viable, also require further technological advances before they can be achieved at scale. Additionally, the logistics of operating data centres in space, from cooling and power supply to maintenance and data transmission, present engineering hurdles that will take years to overcome.
The brokerage firm J.P. Morgan has provided a clear timeline for how it expects SpaceX’s strategy to unfold. Beyond 2029, we expect SpaceX to pivot to orbital compute for incremental capacity additions, while continuing to operate and maintain its terrestrial compute clusters, the firm stated in its analysis. This phased approach suggests that the company envisions orbital infrastructure as a complement to, rather than a replacement for, its terrestrial operations, at least during the coming decade.
For investors and industry observers, this distinction carries important implications. The revenue streams from terrestrial compute contracts are already substantial and growing, providing a solid financial foundation for the company’s more speculative long-term projects. At the same time, the continued investment in ground-based infrastructure ensures that SpaceX remains competitive in an increasingly crowded AI services market, where established players and new entrants alike are racing to secure computational resources.
What makes SpaceX’s position unique is its ability to leverage expertise from its core space operations in its terrestrial AI ventures. The company’s experience with complex systems, rigorous testing protocols, and efficient manufacturing processes translates effectively into the data centre space. There is also a symbiotic relationship between the two tracks of the business, with advances in terrestrial computing potentially informing and accelerating the development of orbital capabilities.
