If you thought the AI infrastructure race was already intense, wait until it leaves the planet. Google and SpaceX are reportedly in early-stage talks to put data centers into orbit, and the implications for the startup ecosystem are far bigger than just another cloud computing partnership.
The Wall Street Journal broke the news that the two companies are exploring orbital data centers — a concept that sounds like science fiction but is increasingly being treated as a serious business proposition. SpaceX, which acquired xAI earlier this year, has been positioning orbital compute as a natural extension of its Starlink network, while Google brings its decades of expertise in large-scale data center operations to the table. A potential deal would come ahead of SpaceX’s highly anticipated IPO, where the company is expected to pitch orbital infrastructure as a key growth driver.
Why Space Makes Sense for AI Compute
The AI boom has created an insatiable demand for compute power, and terrestrial data centers are running into real-world constraints. Land acquisition faces local opposition. Power grids in many regions are strained. Water for cooling is increasingly scarce. And the physical footprint of new data centers is growing as chip densities increase.
Orbital data centers sidestep all of those problems. Once in orbit, they can run on solar power, radiate heat directly into space (much more efficient than ground-based cooling), and operate without the permitting battles that slow down terrestrial builds. Elon Musk has been vocal about the economics, arguing that orbital operations could actually be cheaper in the long run once launch costs continue to fall — and with Starship driving down per-kilogram costs, that calculus shifts faster every quarter.
Google’s Project Suncatcher, announced late last year, already has prototype satellites targeting a 2027 launch. The company has also been talking to other rocket-launch providers beyond SpaceX, suggesting this isn’t a single-vendor bet but a broader strategic push.
The Economic Reality Check
Before anyone gets too carried away, the near-term economics are brutal. As TechCrunch has previously reported, the cost of building and launching a data center into orbit remains dramatically higher than building one on the ground, even accounting for operational savings. Satellite construction, radiation hardening, launch expenses, and the complexity of in-orbit maintenance create a cost structure that today’s terrestrial operators simply don’t face.
But the key phrase there is “today’s.” Launch costs have fallen more than 90% over the past two decades, and Starship promises to push that further. If orbital data centers become viable within the next five to ten years, the companies that built the infrastructure, software, and business models early will have an insurmountable advantage — much like AWS, Azure, and Google Cloud built today after investing in cloud infrastructure when it was still unprofitable.
What This Means for Startups
For founders, this isn’t just a Big Tech story. An orbital compute layer creates entirely new categories of opportunity:
Edge computing gets a new ceiling. If low-earth orbit becomes a viable place to run inference workloads, latency drops dramatically for global applications. A startup building real-time services — from autonomous logistics to global financial trading — could leapfrog terrestrial infrastructure entirely by designing for orbital compute from day one.
Space tech startups have a real commercial customer. The space industry has long struggled to find genuine commercial demand beyond communications and Earth observation. Orbital data centers would be a massive customer for launch services, satellite manufacturing, in-space servicing, and radiation-hardened electronics.
Software-defined space infrastructure. The companies that build the orchestration layers, failover systems, and developer tools to treat orbital compute as part of a unified cloud will be the next generation of infrastructure unicorns.
The Takeaway
Google and SpaceX’s orbital data center talks signal something bigger than a single deal. They represent the recognition that AI compute demand will outstrip what Earth-based infrastructure can economically provide. The startups that recognize this shift early — and build for a world where compute lives both on the ground and in orbit — will be positioned to ride the next infrastructure wave. The cloud moved to the edge. Now it’s moving to space.
Based on reporting by TechCrunch.