Market
InvestAI, and what a gigafactory needs underneath it
The European Commission's InvestAI initiative directs public money, and mobilises private money, toward AI capacity in Europe – including sites at gigafactory scale. The money is the easy part. What a site must bring is a set of physical properties that take years to assemble and cannot be bought in a funding round.
Europe has committed money to AI capacity. The constraint was never the money, and the shortlist of sites that can take it is short.
What the money is for
InvestAI is aimed at compute capacity on European soil: large training sites and the infrastructure beneath them. The stated logic is that a continent unable to train models on its own ground has outsourced a strategic capability. That logic points at substations, cooling and buildings before it points at accelerators.
The bottleneck was never capital
A site at that scale needs a grid connection in the hundreds of megawatts, a cooling source that does not consume drinking water, land with no neighbour close enough to object, and a legal position that survives due diligence. Each of those has a lead time measured in years. Capital arrives faster than any of them, which is why announced capacity and connected capacity are different numbers in every European market.
What it means for a smaller site
Not every project is a gigafactory, and pretending otherwise is how a development gets priced wrong. A one-to-three megawatt building in an existing shell serves regional customers who cannot wait for a campus and do not need one. A fifty megawatt phase serves enterprise and sovereign workloads. They are different products on the same standards, and an honest pipeline says which is which, and when.
Sources
- 01European Commission – InvestAI and the AI gigafactories programme
- 02⚠️ Programme figures move between announcements. Quote the Commission's current page, never a press summary.
