New York Governor Kathy Hochul thinks AI should be ‘less evil’

Near term this reads as a regulatory-template risk more than an earnings event. META is the cleanest loser because any state-backed age-verification stack raises friction in the highest-velocity part of the ad funnel: more logins, more drop-off, weaker targeting, and more compliance cost per impression. The second-order effect is that authenticated ecosystems with existing identity rails and distribution breadth are better insulated, so GOOGL is comparatively better positioned than pure social peers if “verified internet” becomes a wider norm; smaller publishers and open-web ad intermediaries would take the real margin hit.

Over 1-3 months, the catalyst is whether this becomes a copy-paste policy model in other states and whether platform-specific settlement language survives scrutiny unchanged. The main reversal risk is litigation or implementation narrowing: if courts force a lighter-touch mechanism, the multiple compression in META should unwind quickly. For AMZN, the data-center politics matter less for current revenue than for 6-18 month capex ROI: local permitting and community-benefit demands raise the hurdle rate for every new build, but the scale players with power procurement and balance-sheet flexibility will be the ones who keep taking share.

Consensus is probably overpricing immediate revenue damage and underpricing the precedent effect. This is less about one settlement and more about a framework for higher compliance, slower product iteration, and more expensive customer acquisition across the internet stack. The market should stay focused on whether the policy becomes a de facto identity layer for the web; if not, the selloff in the most exposed names is likely to prove tactical rather than structural.