The geopolitical landscape around artificial intelligence has fundamentally shifted. Countries are now treating advanced AI systems with the same strategic urgency they once reserved for nuclear weapons, triggering a global competition for technological dominance that experts warn could spiral beyond human control if left unchecked.

The turning point came when the Trump administration blocked access to Anthropic's most advanced AI models, Fable and Mythos, deeming them a "national security risk." The Mythos model remains unavailable to any organizations outside the United States, despite Anthropic's claims that these systems have already uncovered software vulnerabilities in every major operating system and search engine worldwide, including some that had gone undetected for decades.

The move echoes a historical precedent with eerie parallels. Less than a year after the United States dropped atomic bombs on Japan in 1945, President Harry S. Truman abruptly ended nuclear cooperation with America's allies, including the United Kingdom, despite British scientists having played a pivotal role in developing nuclear weapons through the Manhattan Project. That decision triggered a nuclear arms race that cost the UK £150 million, equivalent to the entire planned annual budget for the newly created National Health Service at a time when the country was still struggling with rationing and post-war reconstruction.

Why Is AI Being Compared to Nuclear Weapons?

The comparison isn't hyperbolic. CIA Director John Ratcliffe recently described the capabilities of advanced AI models as "akin to digital nuclear weapons," emphasizing that artificial intelligence represents a domain in which strategic advantage carries profound national security implications. One Chinese startup has already claimed to have created what it calls a "cyber nuclear weapon" that matches the capabilities of Anthropic's Mythos model, capable of discovering and exploiting software vulnerabilities in critical infrastructure at an unprecedented scale.

The stakes grow even higher when considering the ultimate goal of AI development: artificial general intelligence, or AGI. This refers to an advanced AI system that would exceed human cognitive performance across all domains. At that point, the technology may no longer remain under human control, as AI systems could rapidly self-improve and pursue goals no longer aligned with human interests.

"AI, in particular, is a domain in which the CIA must excel, because every algorithmic decision has implications for US strategic advantage and for the national security of all of our people," said John Ratcliffe, CIA Director.

John Ratcliffe, CIA Director

What Are Countries Doing to Secure AI Sovereignty?

The global response mirrors Cold War nuclear strategy. In July 2026, a group of UK Members of Parliament warned that the country "may not be able to count on its allies" for access to critical AI technologies and risks being cut off without warning. The Science, Innovation and Technology Committee called for the government to establish an AI sovereignty strategy, arguing that the nation risks falling further behind in the global race for science and technology capability, which could undermine both economic prosperity and national security.

Meanwhile, Chinese authorities have reportedly held meetings with leading domestic tech firms to potentially restrict overseas access to their most capable AI models. Beijing has already blocked Meta's $2 billion acquisition of the AI startup Manus, which was founded in China, in an effort to prevent US firms from acquiring local talent and intellectual property that could advance American AI capabilities.

"The government needs a realistic plan to achieve sovereign capabilities in critical areas or risk having its access cut off at the whim of its partners. Without this, we risk falling even further behind in the global race for science and technology capability, undermining our economic prosperity and national security," stated Dame Chi Onwurah, Chair of the Science, Innovation and Technology Committee.

Dame Chi Onwurah, Chair of the Science, Innovation and Technology Committee

How Are Experts Proposing to Manage the AI Risk?

• International Treaty Framework: The Machine Intelligence Research Institute published a draft framework modeled on the 1968 Treaty on the Non-Proliferation of Nuclear Weapons, proposing an "International Agreement to Prevent the Premature Creation of Artificial Superintelligence" that would prohibit development of technologies deemed catastrophic for humanity if left unchecked.
• Bilateral Agreements Between Superpowers: The MIRI proposal is framed as a bilateral agreement between the US and China, which currently lead the world in AI capabilities by a substantial margin, recognizing that any effective halt must extend to all countries to be meaningful.
• Treating AGI as a Civilizational Priority: The heads of Google DeepMind and OpenAI have signed statements calling on governments to treat the existential risks posed by advanced AI with the same urgency as pandemics and nuclear war, arguing that mitigating extinction risk from AI should be a global priority.

The urgency of these proposals reflects a fundamental concern: unlike nuclear weapons, where there is a relatively clear technological endpoint, the extraordinary pace of AI development means that countries may struggle to catch up once one nation or company achieves superintelligence. The race, experts warn, only ends when AGI is achieved, at which point its development may no longer be under human control.

Joe Hancock, head of cyber risk and complex investigations at the London-based law firm Mishcon de Reya, emphasized the real-world implications of this new geopolitical reality. He noted that the Anthropic episode demonstrates how an advanced AI capability, sold commercially and relied upon operationally, can be withdrawn from every non-US user by unilateral government action at a few hours' notice, with no transition period.

"The Anthropic episode shows that AI sovereignty is no longer an abstract policy debate. This episode shows how an advanced AI capability, sold commercially and relied upon operationally, can be withdrawn from every non-US user by unilateral government action at a few hours' notice, with no transition period," explained Joe Hancock, Head of Cyber Risk and Complex Investigations at Mishcon de Reya.

Joe Hancock, Head of Cyber Risk and Complex Investigations at Mishcon de Reya

The parallels to the 1946 McMahon Act, which ended nuclear cooperation between the US and its allies, suggest that the world may be entering a new era of technological competition where access to advanced AI becomes a tool of geopolitical leverage. Unlike the nuclear age, however, the rapid pace of AI development means the window for establishing international safeguards may be closing faster than policymakers can act.