Cory Doctorow says sovereign AI misses the real platform risk

Cory Doctorow, the writer and technology critic who popularized the term “enshittification,” has argued that the push for “sovereign AI” is the wrong priority for countries worried about dependence on the United States. In a post published Wednesday, Doctorow said national efforts to control AI capability matter less than control over the everyday software, data centers and connected equipment that governments and companies already rely on.

The argument lands in a policy debate that has become more urgent as U.S. President Donald Trump weakens parts of the U.S.-led order that have supported American influence in diplomacy, defense, finance and technology. Australian writer John Quiggin framed the issue as a risk for countries that do not want critical technology exposure to a less predictable U.S. government.

Quiggin focused on AI dominance as one possible point of leverage. He suggested that the current spending model at frontier AI labs may not determine the outcome of the category, arguing that lower-cost second movers such as DeepSeek and France’s Mistral could reproduce capabilities roughly a year behind U.S. leaders without the same dependence on large data center buildouts.

Doctorow’s response was more skeptical of AI’s strategic importance. He argued that the case for sovereign AI assumes AI systems are already important enough to justify national industrial policy at the center of digital resilience. In his view, that premise remains unproven.

His comparison is operational rather than theoretical. If the Trump administration cut off foreign access to AI chatbots, Doctorow argued, the effect on most businesses would likely be limited. If the administration ordered Microsoft to switch off Office 365 for foreign users, or pushed John Deere to disable connected tractors, the economic damage would be immediate and measurable.

That distinction matters for technology buyers and policymakers. Enterprise AI tools still sit on top of the cloud, productivity suites, identity systems, telecoms infrastructure, payment rails, e-commerce software and industrial equipment that run daily operations. Doctorow’s point is that countries cannot build resilience by funding national AI models while leaving those deeper dependencies untouched.

Doctorow has been a consistent AI skeptic, including in his recent book, The Reverse Centaur’s Guide to Life After AI. The title refers to an image of humans being directed by a machine system, a framing that fits his broader criticism of companies promoting AI as a general solution while its commercial returns and reliability remain contested.

The practical alternative he outlined is less fashionable than sovereign AI: build and operate domestic or trusted platforms for routine administrative, telecoms and e-commerce workloads, then move public-sector, corporate and household data onto them. He described that as the real digital challenge facing countries concerned about U.S. leverage.

The broader warning is aimed at Europe and other U.S. allies. American technology companies, according to the argument, have shown little willingness to resist pressure from the U.S. government unless political incentives change. If governments want bargaining power, Doctorow’s case is that they should reduce exposure to existing platform chokepoints before treating AI autonomy as the central problem.

This story draws on original reporting from The Register.