Is the Age of Purely Private-Sector AI Coming to an End?

In this insightful piece for CIGI, author S. Yash Kalash explores how the “nationalization” of artificial intelligence is already emerging through equity stakes, public subsidies, government procurement, export controls, compute access and sovereign investment, rather than outright government takeovers. The United States, India, Europe, the UK, Gulf states and China are pursuing different approaches, but share a growing interest in ensuring that critical AI infrastructure does not remain entirely dependent on private or foreign actors.

The author argues that limited public ownership could also help address the growing legitimacy crisis facing major AI companies. Small, non-controlling government stakes could give citizens a share in AI-driven economic growth while providing firms with greater political and social legitimacy. Returns could support public compute, worker retraining, schools, small businesses and infrastructure, creating a broader public-benefit compact rather than simply generating government revenue.

However, public ownership creates significant risks if governments become simultaneously shareholders, customers, regulators, and competition authorities. Here the piece stresses that state participation must not shield AI companies from antitrust enforcement, liability, safety regulation or competitive procurement. Any government stake should therefore be held through a ring-fenced public or sovereign AI fund, with limited voting rights and strong institutional firewalls preventing political control over AI systems or regulatory decisions.

The article continues, arguing that the emerging model should resemble a passive sovereign wealth holding combined with a public-benefit agreement, rather than traditional nationalization. Public support for AI should generate reciprocal benefits such as affordable compute, research access, safety testing, worker retraining and critical infrastructure investment, while avoiding taxpayer-funded bailouts of unprofitable firms. The central challenge, he concludes, is ensuring that public participation expands accountability and shares AI’s economic upside without turning governments into political gatekeepers, corporate protectors or buyers of last resort…

CONTINUE READING…