Google and SpaceX Push Orbital Data Centers Forward

According to Bisnow, Google and SpaceX are accelerating efforts to move AI infrastructure beyond Earth, turning orbital data centers from a fringe concept into a serious frontier for the tech and infrastructure industries. According to The Wall Street Journal, the companies are in talks over a launch agreement that would see SpaceX carry Google computing units into low Earth orbit as part of Google’s broader Project Suncatcher initiative.

The discussions come as hyperscalers, AI firms, and private space companies scramble to secure enough compute capacity and power to support next-generation artificial intelligence models. With terrestrial data centers facing mounting grid constraints, water concerns, and permitting delays, orbital infrastructure is increasingly being pitched as a long-term workaround powered by uninterrupted solar energy.

How Orbital Data Centers Became Viable

The idea of placing computing infrastructure in space has circulated for years. However, the AI boom has changed the economics behind the concept. Massive AI training clusters require enormous amounts of electricity. Hyperscale campuses now demand gigawatts of power across major US markets.

Google entered the race in November 2025 with Project Suncatcher. The initiative aims to deploy AI computing nodes across interconnected satellites in low Earth orbit. According to the WSJ, Google plans to launch two prototype satellites by early 2027. The satellites will carry the company’s proprietary AI chips.

Google CEO Sundar Pichai has framed orbital computing as the next step in digital infrastructure. He no longer treats the concept like science fiction. That shift matters for the broader industry. Major cloud operators now view space-based AI infrastructure as strategically important. Still, commercial deployment likely remains years away.

The Details Behind the Google-SpaceX Talks

The reported discussions center on launch logistics, one of the biggest bottlenecks facing orbital data centers today. SpaceX remains the dominant commercial launch provider globally, giving it a major advantage as companies look for reliable heavy-lift capacity.

The company is simultaneously building its own space-based AI ambitions. Earlier this year, SpaceX filed plans with the Federal Communications Commission to deploy as many as 1M solar-powered satellites designed to function as distributed AI data centers. Elon Musk later merged SpaceX with xAI, explicitly tying the rocket business to future AI infrastructure deployment.

SpaceX has also begun integrating orbital computing into customer relationships. Earlier this month, Anthropic reportedly discussed future space-based AI capacity with SpaceX while securing 300 MW of compute at the company’s Colossus data center campus in Memphis, Tennessee, according to the WSJ.

That overlap between terrestrial and orbital infrastructure is becoming a defining feature of the sector. Most firms pursuing orbital computing still rely heavily on Earth-based campuses while developing long-term deployment strategies for space.

A Crowded Race for Space-Based AI Computing

The orbital data center market is quickly attracting startups, defense-adjacent firms, and billionaire-backed space companies. This week, Cowboy Space — formerly known as Aetherflux — closed a $275M funding round that reportedly valued the company near $2B.

The company’s strategy goes beyond building data centers in orbit. Cowboy Space is developing its own integrated launch system, embedding AI computing infrastructure directly into a rocket’s second stage to reduce deployment complexity and dependence on third-party launch providers.

Founder Baiju Bhatt told TechCrunch the company expects its first launch by late 2028. The strategy reflects growing concerns about launch scarcity and payload competition. Meanwhile, several aerospace firms already use Nvidia processors for space-based AI workloads, as orbital computing infrastructure gains commercial traction.

Jeff Bezos-backed Blue Origin is pursuing a similarly ambitious strategy through Project Sunrise, a proposed network of 51,600 satellites designed to operate as a unified orbital computing platform. But the project hit a regulatory hurdle this week after NASA filed objections tied to potential interference with manned missions and scientific satellite operations, according to Data Center Dynamics.

Even so, NASA’s response may ultimately validate the sector’s credibility. Rather than dismissing orbital data centers outright, the agency indicated a willingness to work with Blue Origin on resolving operational concerns.

Why Orbital Data Centers Matter for CRE and Infrastructure

The rapid emergence of orbital computing reflects growing stress across traditional data center markets. AI workloads are reshaping power demand forecasts nationwide, with utilities and grid operators struggling to keep pace with hyperscale expansion.

According to CBRE’s 2025 North America Data Center Trends report, vacancy rates across primary data center markets fell below 3% as AI demand accelerated leasing activity and power procurement competition. In several Tier I markets, developers are now waiting years for utility interconnections.

Orbital data centers aim to bypass many of those constraints entirely. Solar-powered infrastructure in orbit could theoretically operate continuously without relying on terrestrial grids, water-intensive cooling systems, or increasingly difficult zoning approvals.

Still, the economics remain highly speculative. Launch costs, maintenance challenges, radiation exposure, and data transmission latency all represent major barriers to commercialization. Regulatory scrutiny is also likely to intensify as satellite volumes increase and concerns about orbital congestion grow.

For commercial real estate players, the trend may not replace terrestrial campuses anytime soon, but it could influence long-term infrastructure planning. Developers, utilities, and institutional investors are already recalibrating around AI’s energy demands, and orbital computing adds another layer to that conversation.

What’s Next for Orbital Infrastructure

The next two years will likely determine whether orbital data centers remain experimental or begin transitioning toward early commercialization. Prototype launches from Google and other firms are expected by 2027, while startup deployments could follow shortly after.

The sector’s progress will depend heavily on launch economics and regulatory coordination. Companies capable of controlling both launch systems and computing infrastructure may hold an early advantage, particularly as payload competition intensifies.

Meanwhile, the broader AI boom continues to create urgency around scalable infrastructure solutions. Whether orbital data centers become mainstream or remain niche, the influx of capital and participation from companies like Google, SpaceX, Blue Origin, and Anthropic suggests the industry no longer views space-based computing as theoretical.