What matters now

Assuming anything like this scale is realised, there will certainly be a lot more traffic in space. However, space really is vast, and even in the orbits selected there should be more than enough space for the satellites themselves.  

There will be a problem with debris though. The satellites will be positioned in orbits between 500 and 2000km, and particularly between 750 km and 1,000 km the orbital space is very polluted, with millions of pieces of small, fast-moving debris (the product of decades of launches, fragmentation events and anti-satellite tests) many of which are difficult or impossible to track. 

A constellation at this scale will require continuous awareness and frequent small manoeuvres. SpaceX will be confident in their own operations; but those manoeuvres will have knock-on effects for other operators, who are likely to have less fuel margin or less sophisticated tracking capability.

Operators in this environment will therefore be very happy to be getting excellent quality data as to their environment, and automated ways of testing out and then executing their manoeuvres to ensure they avoid the debris without affecting the constellation. 

A compute shift – and an industrial opportunity

Life is more than getting out of the way of each other however, and one of the promises of orbital data centres is the potential for lower-cost, and potentially lower emission compute at scale. 

If it proves achievable, it raises a straightforward question for UK public services: how would we design services differently if compute were materially cheaper and more available?

There are also practical engineering questions. Satellites are often designed for operational lifetimes of seven years or more, while AI hardware can evolve on a much shorter cycle. Replacing entire spacecraft to keep pace with processor development would be expensive and inefficient.

A more sustainable model could involve modular design and in-orbit servicing, allowing high-value components such as processors and memory to be upgraded while retaining longer-life systems such as power, structure and communications. The UK’s industrial strategy specifically cites in-orbit servicing and manufacturing as a strategic goal, and the UK has a strong industrial base in this area. If modular upgrade models gain traction, this could represent a meaningful opportunity.

Planning ahead

Orbital AI data centres, if realised at scale, would both increase the complexity of space operations and create new opportunities. Navigating that balance will require clear digital strategy, credible regulatory approaches, and robust space domain awareness.

If SpaceX does indeed deploy a million satellite constellation, it may not “change everything”. But it probably will change enough to matter. Space will become busier. Compute may become cheaper. Regulation will become more complex. And the countries that think clearly about those shifts early will be better placed than those that do not. These are the kinds of long-horizon infrastructure questions we spend our time working on at The PSC.

If you want to talk about space sustainability, get in touch with Phil at Phil.Buckley@thepsc.co.uk