If we hadn't already invested attention in Mistral AI's fundraising, would we really conclude it's solid proof for the AI Act's viability? Their reported 21 billion euro valuation is a sunk cost for investors, reflecting a moment in time, not a guarantee the whole framework works. You see it clearly when you look at their actual capital allocation strategy: they're pouring money into servers and hiring smart people, not regulatory compliance. It's like saying my son doing well in maths proves the entire school system is perfect, ignoring the teacher he pays extra for.