Trump Signal Index
On September 14, Trump posted to Truth Social that a “disgusting conspiracy” is targeting AI and data centers — one that “only China would celebrate.”
“Whoever wins AI, wins,” he wrote. “We are ahead of China and every other country, and we will stay that way.” Then the sentence that moves this story into a different category: “Conspiracists, traitors, betrayers, leakers — be careful.”
From Dismissal to Threat in Three Days
The trajectory has been fast. On September 10, asked whether AI could destroy humanity, Trump said he was “not worried at all” — only about losing to China. In Dublin on the 13th, he called the slowdown advocates “negative forces.” Now they are a conspiracy, and its members are traitors who should be careful.
What changed in between is who was talking. The warnings that began with resigning safety researchers at Anthropic and Google have moved up the hierarchy: Anthropic CEO Dario Amodei is now among the industry leaders publicly arguing for slowing development. The dissent stopped being an employee-retention story and became a policy demand from the people who own the labs.
Trump’s response to Amodei specifically: he is “pretending to be a perfect little angel.” And then the part with legal weight — the administration has “stopped people like Amodei from doing bad things or potentially bad things and will continue to,” because “we already have enormous criminal and regulatory power over these companies.”
A president publicly reminding a named CEO of the state’s criminal power, in the context of that CEO’s policy disagreement, is the mechanism to price — separate from anything about AI. It converts regulatory authority from a rulebook into a loyalty instrument: the companies that align (the labs staffing federal commissions, winning defense contracts) operate under one regime, and the companies that dissent operate under implied investigation. This is the same architecture this site has tracked all year — the super PAC that is “my money”, the Fed pressured by trade ultimatums — now extended to the most valuable industry in the country. And his stated alternative to AI regulation completes it: “The only control or safeguard AI needs is a strong, smart, high-IQ president.” Not law, not institutions. Him.
The Substance Being Buried
Underneath the invective, an actual policy question is being closed without being argued. The slowdown advocates’ case — made by OpenAI’s chief scientist in writing — is that frontier systems grow rather than being designed, are not understood by their builders, and need safeguards before further scaling. The advocates’ proposed mechanism is government coordination, domestic and international.
Trump’s position forecloses it twice over: the race with China makes any pause unilateral disarmament, and the “conspiracy” framing makes advocating one disloyal. Whatever the object-level truth about AI risk, the US now has no institutional channel through which a frontier-lab warning could produce a policy response. The safeguard is the president’s IQ, by his own account.
There is an irony worth registering: in the same post that rejects AI safeguards as unnecessary, Trump claims already-existing, actively-used power to stop AI companies from “doing bad things.” Control exists — it is simply reserved for dissent rather than danger.
What It Means for Markets
The buildout floor stands — nothing here slows a data center or a chip order, and the race framing still overrides local opposition. What changed is the introduction of company-specific political risk within the winners’ circle.
The new variable is alignment risk: AI companies now face materially different regulatory exposure depending on their posture toward the administration, and “enormous criminal and regulatory power” invoked against a named CEO is the kind of statement that shows up later in depositions, discovery, and — for investors — in how private-market valuations of dissenting labs get discounted against compliant ones. Watch three things: whether any concrete regulatory or investigative action follows against Anthropic (which would confirm the threat as operational, not rhetorical); whether other lab CEOs go quiet (silence after this post is itself the signal that the threat worked); and whether safety-motivated departures accelerate into a talent migration — the destination of that talent, domestic or foreign, matters for the “one year ahead” claim Trump’s entire framework rests on. The tail nobody prices remains the same as last week, now with an addendum: if a frontier system fails visibly, the record will show the government called its warners traitors, and the regulatory whiplash will be proportional to that record.
The Post
AI safety concerns are a ‘disgusting conspiracy’ that ‘only China would celebrate’
The Threat
‘Conspiracists, traitors, betrayers, leakers — be careful’
The Target
Amodei named: ‘pretending to be a perfect little angel’
The Power Claim
‘Enormous criminal and regulatory power over these companies’ — already used
The Safeguard
‘The only control AI needs is a strong, smart, high-IQ president’
The Context
Researcher resignations at Anthropic and Google; CEOs now urging a slowdown