Project Suncatcher: space data centers 2.5–3x pricier, BCG says

SpaceX's launch of a prototype satellite for Alphabet's Project Suncatcher marks a concrete step in testing whether AI chips can operate in low earth orbit, where they could draw on nearly continuous solar power. Alphabet's longer-term ambition is to run entire constellations of data center satellites, but the body notes that processors must first prove they can work properly and sustainably in space before any deployment at scale.

The economic case remains unsettled. Boston Consulting Group estimates that satellite-based data centers will be 2.5 to 3 times more expensive than Earth-based ones on a processing-power basis, even though orbital facilities would avoid electricity costs, with that gap potentially shrinking as launch costs fall. Meanwhile, neoclouds have gained head starts in Earth-based infrastructure, and hyperscalers are paying premiums for compute now, as shown by Meta's deal for a nuclear-powered data center campus in Ohio that will not be completed until 2034.

Even if orbital data centers prove technically and commercially viable, they are unlikely to make Earth-based data centers obsolete, similar to how air freight did not eliminate freight railroads or tractor-trailers. The test is whether the cost gap narrows and whether the hardware can operate reliably in space, which matters most to neocloud operators and hyperscalers planning long-term capacity.