The Tech Titans Just Killed AI Regulation — And No One Is Watching

The Quiet Coup

What happened this week in Washington reads like a masterclass in regulatory capture. Three of the most powerful figures in technology — Jensen Huang at NVIDIA, Mark Zuckerberg at Meta, and Elon Musk at SpaceX — separately sat down with Donald Trump over recent weeks and made the same case: do not create the independent AI regulatory body that experts across the industry had been discussing. Trump listened. The body is now dead.

The Wall Street Journal reported the story on September 17th, but the implications have barely registered outside specialized policy circles. That needs to change.

Who Proposed This, and Why It Would Have Mattered

The regulatory body was not some government bureaucrat’s fantasy. It originated from Demis Hassabis, Google’s chief scientist, who proposed a private-led AI oversight and standards organization modeled on the Financial Industry Regulatory Authority — the FINRA that regulates U.S. securities firms. The idea was straightforward: create an independent body with authority to set safety standards, audit development practices, and enforce compliance across the industry, rather than relying on voluntary commitments that companies could walk away from.

This was significant because AI development is moving at a pace that outstrips legislative processes. Congress cannot write code. It cannot audit model weights. A dedicated regulatory body with technical expertise and enforcement teeth was the most realistic mechanism for imposing even modest guardrails on systems whose capabilities are compounding exponentially.

The Real Motive

Here is what the sources tell us, and it is the part that should alarm anyone paying attention: Huang, Zuckerberg, and Musk did not oppose the regulatory body on ideological grounds alone. They argued, according to insider accounts, that such a body would entrench the power of just three companies — OpenAI, Anthropic, and Google — at the expense of everyone else.

That argument is a perfect mirror of the one monopolies have made against every major regulation in American history. It is also, on its face, plausible. The three companies that would sit at the center of any regulatory framework are the three that already dominate training compute, top-tier talent, and distribution channels. A regulatory body would inevitably elevate the incumbents — they have the resources to comply, the lobbyists to shape the rules, and the data to satisfy auditing requirements.

But the reverse is also true. The very concentration of power that these CEOs warn about is exactly what a regulatory body is designed to check. By killing the body, they preserve a system in which the companies with the most GPUs already dictate the pace, direction, and safety standards of the most consequential technology in human history.

The Trump Calculation

Trump’s response has been unmistakable. He took to Truth Social to declare that “a morbid conspiracy is being orchestrated around AI and data centers” and that “only China benefits from this.” The framing is telling: regulation is not a safety measure, it is an act of economic sabotage, and the only rational alternative is unfettered acceleration.

This is not a novel position for Trump, but the convergence of a Republican president with three tech CEOs who share his anti-regulatory instinct is historically unusual. The White House is internally divided — Deputy Chief of Staff Suzy Wiles, Treasury Secretary Scott Bessent, and National Cyber Director Sean Cairncross have all pushed for stronger AI oversight, while White House Science and Technology Advisor David Sacks has counseled minimal regulation. Sacks won. The administration will pursue what the Journal describes as a “light-touch” approach.

The Hypocrisy Problem

The most striking detail in the reporting is Elon Musk’s contradictory posture. He publicly agreed with Dario Amodei, Anthropic’s CEO, that the industry needs to slow down. He called Amodei “right.” But SpaceX’s own AI division has never published a concrete plan to actually reduce development velocity. When questioned, Musk’s companies continue building at whatever pace the market demands. He can endorse caution rhetorically while benefiting operationally from the absence of constraints.

Zuckerberg’s position is more consistent, if no less self-serving. He has argued that individual companies can and should self-regulate their development speed without external oversight. It is a position that requires faith in the very institutions that have spent decades demonstrating a systemic reluctance to prioritize safety over market share.

Huang occupies the most interesting intellectual ground. At a Salesforce event, he stated that safety is a priority but also an engineering problem — one that can be solved simultaneously with speed. This is a coherent position in theory. In practice, it amounts to the claim that the companies best positioned to solve the safety problem are the same ones that profit most from solving it slowly.

Who Actually Wins

The losers are immediate and clear: anyone who believed that independent oversight could impose meaningful constraints on AI development before the technology outpaces our ability to control it. The regulatory body would have created a formal channel for dissenting voices within the industry — scientists like Hassabis and Amodei who warned about concentration of power and unsafe deployment — to translate their concerns into enforceable standards.

The winners are the three companies that now face no structural check on their trajectory. OpenAI, Anthropic, and Google’s DeepMind will continue competing for dominance without an independent auditor asking whether their race to the bottom on safety is producing outcomes anyone besides shareholders benefit from.

There is also a broader loser that deserves naming: the public interest in democratic oversight of technologies that are already reshaping labor markets, information ecosystems, and geopolitical dynamics. When three private actors persuade a president to abandon regulatory architecture, the conversation that should have been public becomes a boardroom decision.

What Happens Next

The absence of a regulatory body does not mean the question goes away. Congress may eventually act. State-level legislation could fill the vacuum. International partners, particularly in the EU, will continue developing frameworks that may apply extraterritorially to U.S. companies doing business in their markets.

But the immediate consequence is unambiguous: the most ambitious attempt to create independent AI oversight in the United States has been defeated by the very companies it was designed to regulate, using the argument that regulation would help those same companies entrench their dominance. It is the oldest trick in the monopoly playbook, and it worked.

The English-language press has barely covered this story. That is itself a signal. When the people most affected by a decision are not the ones reading about it, something has already gone wrong.