The High Table
The Bug in the Black Box: When AI Labs Become Sovereign Territory
The Saudi bugging of OpenAI isn't espionage — it's a recognition ceremony. Nation-states have quietly acknowledged that frontier AI labs now wield sovereign-level power, and they're applying the old playbooks of statecraft to these new territories. The recording device in the conference room is merely the crudest instrument of a deeper shift: the privatization of strategic capability has created a new class of quasi-sovereign entities that states must penetrate, align with, or control.
By Spencer Vance
October 19, 2026
12:06 PM PDT
The Frame
In January 2025, three weeks into his second term, President Trump signed Executive Order 14179 — the “AI Sovereignty Directive” — classifying frontier model weights above a certain compute threshold as “critical national infrastructure” subject to CFIUS review, export controls, and a new category of protection: “strategic information assets.” The order was drafted in the Situation Room, not the West Wing. Its existence was leaked not by a whistleblower but by a venture capitalist who noticed unusual clause insertions in term sheets for Series C rounds at three major labs. The Saudi bugging case, unsealed in October 2026, is the first public evidence that foreign intelligence services received the memo before the American public did. They understood something Washington’s policy apparatus is still struggling to articulate: the labs have become the new embassies, and their employees the new diplomatic corps — but without the Vienna Convention’s protections.
The Signal
What most...
The High Table - Founding Circle
Exclusive content for High Table members. Invitation required.
Learn More