Why ExoCompute Exists

About 18 months ago, my former cofounder Tom Ray and I sat down to figure out what to build next together. We ran through a bunch of interesting ideas and a few unusual ones, including a crazy one: data centres in space. We didn't go for it, but I created ExoCompute because space is the place and the idea of sticking compute out there is awesome.

Ground space and energy

Building a data centre today costs between $7–12 million per megawatt of IT capacity. AI-optimised facilities regularly exceed $20 million per MW. The US development pipeline alone is now approaching $2.3 trillion (Cushman & Wakefield, 2026).

Cost is only part of the story. Communities are pushing back. From the Netherlands to Northern Virginia to planning committees across the UK and Ireland, residents and local authorities are blocking new data centre developments. They are objecting to noise, power requirements, water consumption, and the strain on local grid infrastructure.

Meanwhile, demand keeps accelerating. AI could represent half of all data centre workloads by 2030. The five largest technology companies spent over $400 billion on infrastructure in 2025 alone, and are expected to increase that by another 75% in 2026. Data centres already consume roughly 1.5% of global electricity, and that share is growing fast (IEA, 2025–2026).

What is ExoCompute?

For now, it's an intel aggregator. Every space launch is reported. Some of those launches will carry compute moving to orbit i.e. edge processing, AI inference, communications infrastructure. ExoCompute scores every mission for its orbital compute relevance and gives you the intelligence layer on top: what's launching, why it matters, and what it signals.

— James Sandoval, Founder, ExoCompute