Let's rethink a term that's quietly holding us back.
"Data center" evokes rows of hard drives quietly storing files - backups, websites, emails from the early internet era. That's still a real need, and those facilities deserve the name. No argument there.
But look at what's driving the explosive new builds, the hyperscale monsters consuming gigawatts and packing tens of thousands of GPUs. These aren't about storing data anymore.
They're factories producing intelligence - AI hubs.
Raw electricity and silicon go in - trained models, reasoning capabilities, synthetic data come out. NVIDIA's Jensen Huang calls them "AI factories," and the metaphor fits perfectly. We take energy, compute, and datasets, then manufacture intelligence at industrial scale.
The orbital concepts taking shape right now make this even clearer. Starcloud, Aetherflux, SpaceX filings, Google's explorations - all focus on unlimited solar power to run nonstop AI training far above Earth's grids. Storage is almost an afterthought. These are pure intelligence production facilities in space.
Imagine entire constellations of these hubs training models the size of civilizations - what becomes possible then?
So why keep calling them "data centers"? The label hides the revolution.
Traditional storage-focused sites can stay "data centers." That's fair.
But the new generation - terrestrial or orbital - needs a name that reflects what they actually do: create AI at scale.
"AI factories" captures the industrial reality.
"AI hubs" feels cleaner for the orbital era - evoking centralized nodes of boundless compute, linked by laser comms, powering a multi-planetary intelligence network.
Either beats clinging to an outdated term.
The words we use shape how boldly we think. Updating it here would force us to see these facilities for what they've become: engines of abundance.
What do you think we should call them - AI factories, AI hubs, or something better? 馃馃殌
