# Why America Will Probably Nationalise AI

The United States stands at a crossroads with artificial intelligence. While the technology has been driven by private enterprise, a growing consensus among policymakers, economists, and national security experts suggests that the federal government may soon take direct control of the most advanced AI systems. This is not a fringe conspiracy theory; it is a logical outcome of three converging pressures: national security, economic stability, and the unique nature of AI itself.

The National Security Imperative

The primary driver for nationalisation is the strategic importance of AI. Unlike previous technologies—such as the internet or smartphones—AI poses a direct threat to national security if controlled by a single corporation or a hostile state. The U.S. government has already classified advanced AI chips and models as dual-use technologies, subject to export controls. But many analysts argue that export controls alone are insufficient.

If a private company develops an AI system capable of designing novel bioweapons, automating cyberattacks, or destabilising financial markets, the government cannot afford to rely on corporate goodwill for oversight. The logic mirrors the nationalisation of nuclear technology during the Manhattan Project: when a technology becomes too dangerous for private hands, the state must step in.

The Economic Stability Argument

The second pressure is economic. AI has the potential to disrupt labour markets on an unprecedented scale. Studies from Goldman Sachs and McKinsey estimate that generative AI could automate up to 300 million jobs globally. In the United States, this could lead to mass unemployment, social unrest, and a collapse in consumer demand.

Private companies have no incentive to manage these negative externalities. Their goal is profit maximisation, not social stability. The government, by contrast, must ensure that the benefits of AI are distributed broadly—or at least that the costs of disruption do not tear society apart. Nationalisation would allow the government to slow or redirect AI deployment, retrain workers, and implement universal basic income funded by AI-generated revenues.

The Unique Nature of AI

AI is not like other technologies. It exhibits what economists call "increasing returns to scale": the more data and compute power a model has, the smarter it becomes. This creates a natural monopoly. In many industries, monopolies are tolerated or regulated. But with AI, the monopoly holder could eventually surpass human intelligence in every domain.

If one company achieves artificial general intelligence (AGI)—an AI that can perform any intellectual task better than a human—it would have unprecedented power. It could manipulate elections, control information flows, and even design new forms of warfare. No democracy can allow such power to remain in private hands. The only responsible option is to bring AGI development under public ownership and democratic control.

Historical Precedents

Nationalisation of critical technologies is not unprecedented. The U.S. government created the Tennessee Valley Authority to manage electricity, the National Aeronautics and Space Administration (NASA) to lead space exploration, and the Defense Advanced Research Projects Agency (DARPA) to develop the internet. In each case, the private sector was either unwilling or unable to bear the risks and long-term investments required.

AI is the next frontier. The cost of training frontier models now exceeds $100 million, and the most advanced chips are produced by a single company, Nvidia. The government already funds much of the basic research through agencies like the National Science Foundation and the Department of Energy. Taking full ownership would be a natural extension of existing policy.

The Likely Timeline

Most experts predict that the U.S. will not nationalise all AI, but rather the most capable systems—those approaching or exceeding human-level intelligence. This could happen within the next five to ten years. The trigger might be a major AI-related disaster, such as a financial crash caused by algorithmic trading or a bioweapon leak from an AI-designed molecule.

Once the political will exists, the mechanism is straightforward. The government could invoke the Defense Production Act to seize AI data centres and model weights, citing national security. Alternatively, it could create a public corporation, similar to Amtrak or the U.S. Postal Service, to develop and operate frontier AI systems. Private companies would still compete in narrow, low-risk applications like customer service chatbots or image generation.

Objections and Counterpoints

Critics argue that nationalisation would stifle innovation. They point to the Soviet Union's failure to develop competitive technology under state control. But the comparison is flawed. The U.S. has a robust tradition of public-private partnerships, not central planning. A nationalised AI agency could license its technology to private firms, creating a competitive ecosystem while maintaining ultimate control.

Others worry about government abuse. An AI controlled by the state could be used for mass surveillance or political repression. This is a legitimate concern, but it is not unique to nationalisation. Private companies already sell surveillance tools to governments. The difference is that a democratically accountable AI agency would have oversight from Congress, the courts, and the public. A private corporation has no such accountability.

Conclusion

The nationalisation of AI in America is not a question of if, but when. The combination of national security risks, economic disruption, and the natural monopoly characteristics of advanced AI make private control untenable in the long run. The government already intervenes in markets for food, drugs, and finance. AI is far more consequential.

The debate should not be about whether nationalisation will happen, but about how to design the institutions that will manage it. The stakes could not be higher. If done poorly, nationalisation could create a surveillance state. If done well, it could ensure that the most powerful technology in human history serves the public interest. The next decade will decide which path America takes.