Silicon Valley’s AI Regulation U-Turn and the Political Chaos Behind It

Silicon Valley spent years blocking AI regulation. Now the industry wants rules. The shift is sudden but understandable.

Executives at leading AI companies see the current patchwork oversight as worse than any Biden-era proposals. They say the Biden administration’s approach was at least structured. Today’s ad hoc restrictions feel chaotic and damaging.

Trump’s voluntary AI executive order dropped on June 2, 2026. It asked companies to submit models for a 30-day government review before release. The order was delayed for months, tangled in internal disputes and opposition from so-called “doomers.”

Silicon Valley’s initial cheerleader for deregulation was President Trump himself. His goal was clear: keep America dominant in AI. “President Trump has clearly and repeatedly articulated his goal: ensure continued American dominance in AI,” said a White House spokesperson.

But the reality became messier. On June 12, the White House slapped export controls on Anthropic’s Mythos 5 and Fable 5 models, fully suspending them worldwide. The move followed a security breach reported by Amazon. Anthropic’s CEO Dario Amodei clashed publicly with officials over AI security concerns, diverging from the White House line despite his company’s participation in the AI Action Plan.

The export controls triggered backlash. Over 175 tech leaders signed an open letter demanding the revocation of restrictions on Anthropic’s models. Industry insiders describe the situation as walking on eggshells amid unclear rules and political fights.

OpenAI faces similar pressure. The Biden administration pressured it to limit the launch of GPT-5.6 Sol to about 20 government-approved partners. OpenAI previewed GPT-5.6 Sol, Terra, and Luna on June 26, confirming a restricted preview at Washington’s request.

OpenAI’s CEO Sam Altman and company officials emphasize this is a short-term measure. They do not want government access to become the permanent norm. They are coordinating with the government on an executive order framework focused on cybersecurity.

The government reviews and approves GPT-5.6 case-by-case during this preview. Even if the preview succeeds, GPT-5.6 may never be globally available on day one. This permissioned rollout contrasts sharply with Silicon Valley’s previous push for open innovation without government interference.

Inside the government, key players discussed the export controls in detail. Treasury Secretary Scott Bessent, National Cyber Director Sean Cairncross, and Commerce Secretary Howard Lutnick all took part. The political dimension of AI regulation is impossible to ignore, said Dean Ball, a former Trump official now at OpenAI. “You can’t ignore the political dimension of this conflict,” he said. “There are missteps on both sides.”

The current standoff highlights Silicon Valley’s political naivety. “They’re politically naive,” said one administration official. “You can’t tell everyone that your product might destroy the world and then not expect the government to be involved.”

Some insiders see the worst outcome unfolding. Dean Ball warns, “The worst possible outcome is the thing that’s happening right now, where Dario just keeps fighting.” Amodei’s confrontations with officials have become a flashpoint in negotiations.

The delay of Trump’s executive order was more than a regulatory setback. It rewarded accelerationists who want rapid AI progress without oversight. Meanwhile, the industry that fought hardest against regulation now begs for a formal framework to bring order to the chaos.

As one AI policy adviser put it, “AI executives who funded Trump’s deregulation push now want a formal framework after chaotic export controls and model restrictions.” Another summed up the mood: “It feels like they’re walking on eggshells a little bit.”

Without clear rules, the AI race risks stumbling into political gridlock and unpredictable government intervention. Silicon Valley’s long gamble against regulation has hit its limits. Now it must navigate the very oversight it once sought to kill.