I've often thought the same thing about the pricing models of centralized AI. Why are they priced the way they are? Is it pure greed, or just inefficiency? - On one hand, the market is correctly assigning big value to AI. $NVDA has a higher market cap than all the companies listed in Germany. - But this market dominance means a tiny group owns the entire stack for models, data, and compute. Pricing here comes from concentration, not regulation. Your point on regulation explains airport FX spreads. It's different for AI pricing as cost friction comes from monopolized infra + closed ecosystems, which create higher per-token, per-request, and per-seat prices. I want the upside of AI to be obvious to everyone, but I don't want the top of the stack extracting all of it. The fix is not nicer gatekeepers. It's open rails that route jobs across many providers and settle instantly. Crypto helps here for the same reason you cite: it reduces the surface area for control at an architectural level.