Artificial intelligence is no longer confined to the data centres of Silicon Valley or the cloud regions dotted around the world. 

It is heading into orbit, hitching a ride on satellites and space stations in a way that could transform defence, climate monitoring, disaster response – and the balance of power itself. Starcloud, Google's Project Suncatcher, SpaceX V3 Starlink satellites, and Axiom Space represent the first wave of the orbital AI race.

When SpaceX undertakes its initial public offering (IPO), as early as next month, its valuation will depend to a large extent on how much credibility its plans to put data centres in space are deemed to be. In space, solar panels can supply constant energy to power the chips running high-capacity AI workloads. But that's only part of the reason why tech companies are scrambling to put data centres in space...

This week on The Business of Tech, I talk to Wellington‑based enterprise architect and AI governance specialist Andreas Hamberger, whose new book Space Mafia explores how quickly “orbital AI” is moving from sci‑fi to infrastructure. 

Drawing on three decades in enterprise tech and a deep background in logic and ethics, Andreas argues that putting high‑capacity AI into space opens up an accountability gap that our laws – and our institutions – are nowhere near ready for.

Heaven or Skynet?

On the upside, orbital AI promises what Hamberger terms a “heaven vector” where satellites analyse live sensor data to spot tsunamis in the Pacific, track major polluters in real time, and give us a planetary‑scale view of climate risk. Done well, it could become an engine of equity, giving every country access to insights that used to belong only to superpowers.

But there’s a darker “Skynet vector”. Space is, in practice, a legal grey zone. When companies start training models and running inference beyond the reach of terrestrial copyright, privacy and weapons laws, who are they accountable to? In Space Mafia, Andreas shows how orbit could become the ultimate jurisdictional escape hatch, a place to crunch stolen data, generate “kill lists”, or run ethically dubious experiments with almost no legal friction.

In our conversation, we dig into four real‑world case studies, from data‑centre constellations through to human‑genome work and defence systems that blend orbital AI with hypersonic weapons. Andreas explains why small countries like New Zealand, one of a handful that has space launch capability thanks to Rocket Lab, are unexpectedly central to this story, how new regulations here and in Europe might bite, and what boards, architects and founders should be doing now to close the accountability gap before it’s too late.

Listen to my full conversation with Andreas Hamberger in episode 150 of The Business of Tech, streaming on iHeartRadio, Spotify, Apple, or wherever you get your podcasts.

Show notes

Space Mafia: The Battle Between an Accountable "Heaven" and an Unfettered "Skynet" in Orbital AI 

Space Mafia - the documentary - Andreas Hamberger

SpaceX and Google Are in Talks to Launch Data Centres in Orbit - Wall Street Journal

Data Centres in Space: A Pipe Dream, or AI’s Next Big Thing? - Wall Street Journal