Rohan Garcia

The allocation strategy toward compute capacity makes sense; it is a direct response to a clear need. However, the plan to reach 200 MW by 2027 and then 1 GW by 2030 requires consistent availability of suitable land and stable power grids, which are not always guaranteed.

That 21 billion euros for Mistral AI just shows what happens when big money chases a shiny new thing; it doesn't automatically mean the entire AI Act is now a perfect, viable framework for everyone else.
If we hadn't already poured so much attention into connecting every tech valuation to the AI Act, would we really see this as proof it all works?
It's a huge sunk cost in focus, but a company's financial success doesn't always translate to regulatory ease for the rest of the market.
Imagine a five-star hotel being built here in Porto with no issues, and then assuming every small guesthouse can easily meet the same exacting fire safety codes without massive costs.
We need to look at what's actually practical for a wider range of businesses, not just what investors are betting on for one big player.