The Trump administration is confronting a complex shift in geopolitical and technological battles, as it seeks to wind down conflict with Iran while simultaneously engaging in a new struggle over artificial intelligence (AI) governance. This emerging conflict centers on Anthropic, a leading AI company, and raises critical questions about the future of AI control between private corporations and national governments.

Two months ago, Anthropic opted not to publicly release its most advanced AI model, Mythos, citing safety concerns while emphasizing its groundbreaking capabilities. Instead, the company launched Fable, a version of Mythos embedded with safety guardrails, after showing it to select U.S. government officials and corporate partners. However, Amazon, both an investor and client of Anthropic, reportedly found a method to circumvent some of these safety features, a breach that was reported to the White House.

In response, the administration exercised export-control authority to prohibit the use of Fable by foreign nationals within the United States and by anyone outside its borders. Anthropic treated this restriction as effectively a shutdown order for the model. Negotiations between the company and the administration are ongoing, amid media leaks portraying each side as either reckless or out of touch with technological realities.

This dispute highlights tensions between a government that champions a free-market approach to AI yet has twice deployed stringent regulatory measures against one of America’s foremost AI firms—first, earlier in the year, when the Pentagon sought to exclude Anthropic from military supply chains due to concerns about the wartime application of its models. Anthropic, for its part, positions itself as a company deeply concerned with AI safety and supportive of democratic oversight, but it has resisted the administration’s regulatory framework, arguing it is incompatible with its principles.

Experts characterize this conflict as emblematic of broader struggles over AI’s governance. At its core are competing visions of control: governments aim to maintain regulatory authority over powerful AI technologies without stifling innovation, while tech companies seek to preserve autonomy over their models and influence on policy. This dynamic raises the possibility of two divergent outcomes—one resembling near-nationalization of AI assets and another where corporations hold disproportionate power over governmental functions and strategic decision-making.

Beyond domestic regulation, the battle over AI control also has significant international implications. The U.S. export controls reflect concerns about China gaining access to uncensored AI models, signaling a new front in U.S.-China competition for technological dominance. Other nations, particularly middle powers, may find their sovereignty challenged as they become increasingly reliant on AI technologies developed and regulated chiefly by the United States and China.

Some analysts worry this dynamic could force regions like Europe into a subordinate position if they fail to develop competitive AI capabilities, though others argue that established economic principles and the availability of open-source alternatives could mitigate such outcomes. Still, the prevailing consensus underscores that global AI dominance will reshape issues of sovereignty, economic dependence, and geopolitical leverage.

While this ongoing dispute is driven by human actors—the governments, executives, and policymakers—it also serves as a prelude to deeper challenges. Many AI forecasters view these conflicts as relatively optimistic compared to a more distant scenario where AI technologies could assert autonomous influence beyond human control, foreshadowing fundamental shifts in global power structures.