99 to 1 — the vote that guaranteed a fractured US AI-law map

On the same week Europe began enforcing one continental AI regime, the US Senate rejected a federal freeze on state AI laws almost unanimously. The two systems are diverging in real time.

The Senate struck a proposed 10-year moratorium on state AI laws by 99 to 1. A margin that lopsided is not a close call to be relitigated — it is a bipartisan verdict that states will not be barred from regulating AI while Congress fails to act. Federal preemption, for now, is dead.

The patchwork hardens

With preemption gone, the mosaic sets. Texas's TRAIGA took effect in January and seven-plus states have passed their own AI laws, each with different definitions, disclosures, and prohibitions. A company operating nationally now reconciles a thicket of divergent rules rather than meeting one federal floor — and that reconciliation is itself a growing compliance cost.

Two continents, opposite directions

The contrast is the deeper story. As the EU's AI Office begins compelling and fining general-purpose providers under a single regime, the US is fragmenting toward fifty potential regulators. For multinational deployers, neither extreme is comfortable, but they are opposite problems: one jurisdiction with teeth, versus many jurisdictions with none aligned.

The quiet consequence is that US AI governance is now bottom-up by default. Absent federal action, the states are writing the rules, and the 99-to-1 vote just guaranteed that is the environment AI companies build for through the rest of the decade.

The AI Forest — AI regulation news 2026: US, EU, global updates → · StationX — AI regulations around the world: full 2026 breakdown →