Geopolitics & Sovereign Positioning

The US Commerce Department has withdrawn a draft regulation that would have restricted exports of artificial intelligence chips to anywhere in the world without US approval, according to an electronic notification posted on a government website, Bloomberg reports. The rule, which had not been publicly released in full, would have created a global permitting regime for advanced AI semiconductors, extending controls far beyond existing restrictions targeting China and other adversarial nations. The withdrawal comes as the Trump administration faces competing pressures: industry concerns over competitiveness and the administrative burden of a global licensing system, versus hawkish voices pushing for tighter restrictions on AI diffusion.

The move follows months of debate over how to balance national security concerns with the economic interests of US chipmakers like Nvidia, which dominates the global AI accelerator market. Export controls remain a central tool in the US strategy to maintain AI superiority, but this withdrawal suggests the administration may be recalibrating toward more targeted measures rather than blanket global restrictions. It also reflects the difficulty of enforcing such controls when allies in Europe and Asia are simultaneously building their own AI chip supply chains and may resist US extraterritorial jurisdiction.

Anthropic is locked in a standoff with the Pentagon over how its AI systems can be used for military purposes, according to The Guardian and Wired. The dispute centers not on whether Anthropic's Claude AI will support defense applications — that question has been settled in favor of collaboration — but on the specifics of how the technology is deployed, particularly in intelligence analysis and operational planning. Software demos and Pentagon records detail how chatbots like Claude could help the Pentagon analyze intelligence and suggest next steps in conflict scenarios, raising questions about human oversight and decision-making authority.

This marks a significant shift from less than a decade ago when Google employees successfully pressured the company to abandon its Project Maven contract, which used AI for drone footage analysis. The current debate reflects how Silicon Valley's relationship with the defense establishment has evolved under the Trump administration, with major AI labs now competing for Pentagon contracts rather than boycotting them. Anthropic's position appears to be about setting boundaries on autonomous decision-making and offensive applications rather than refusing military work outright.

Bloomberg reports that former CIA and intelligence community officers are founding defense technology companies at an accelerating rate, seeking to capitalize on Trump's proposed $1.5 trillion defense budget and an estimated $338 billion market for AI and emerging technologies. Brian Carbaugh, a former CIA officer, now leads Andesite, a Virginia-based data analytics startup, exemplifying the trend of intelligence professionals transitioning from operational roles to the C-suite. The shift reflects the Pentagon's growing preference for working with agile startups that understand both the technology and the operational requirements of intelligence and military missions, rather than relying solely on traditional defense contractors.

This development represents a structural change in how national security capabilities are developed and deployed. Intelligence agencies are increasingly viewing partnerships with private AI companies as essential to maintaining technological advantage, particularly against adversaries like China who are integrating commercial AI advances directly into military and surveillance systems. The trend also raises questions about conflicts of interest, security clearances for private companies, and the revolving door between intelligence agencies and the defense technology sector.

The UK government plans to urge the NHS and Ministry of Defence to prioritize British technology purchases as part of a new industrial strategy that banks on artificial intelligence to drive economic growth, The Guardian reports. Treasury minister Spencer Livermore previewed the strategy ahead of a high-profile speech by Chancellor Rachel Reeves, who is expected to restate her economic agenda on Tuesday even as rocketing oil prices from the Iran crisis threaten fiscal targets and growth projections. The policy represents a form of industrial nationalism aimed at building sovereign AI and technology capabilities while supporting domestic companies.

The timing is precarious: the government is betting on AI productivity gains to offset economic headwinds from the Iran conflict, including energy-driven cost increases across the economy. The strategy also intersects with the UK's broader efforts to position itself as an AI superpower, which have included major datacenter investments and regulatory initiatives. However, questions remain about whether the NHS and MoD have the procurement flexibility to favor domestic suppliers without running afoul of trade commitments or sacrificing cost-effectiveness and technological performance.