While governments were busy designing tariff walls and debating the revival of 20th-century industrial policy, something extraordinary happened to global trade: artificial intelligence quietly became its fastest-growing engine. AI-related trade grew at nearly 40 percent in 2025, according to McKinsey Global Institute data, against a global trade average of 6.5 percent. In the noise of the tariff wars, the most significant economic transformation of the decade was unfolding largely unnoticed. This is not merely a data point. It is a verdict on the mismatch between the trade policies that governments are deploying and the economic reality that trade is now expressing. Policymakers are fighting the last war. The battlefield has moved. Consider what AI-related trade actually encompasses. It includes the chips that train and run AI models, a category dominated by a handful of companies and constrained by a small number of geographies. It includes the server infrastructure, the networking equipment, and the specialized cooling systems. It includes the software services built on top of these systems, the consulting and integration work, and the data itself, now recognized as a tradeable commodity with distinct geopolitical value. Together, these flows constitute a new layer of the global economy, one that did not meaningfully exist a decade ago and now accounts for a growing share of the world’s most valuable commerce. Three things must change if governments are to remain relevant economic actors in this environment. First, industrial policy must be reoriented away from protecting legacy sectors and toward building the capabilities that AI-era trade demands: talent pipelines, semiconductor fabrication, data infrastructure, and the regulatory frameworks that govern how AI systems are deployed. The countries that get this right in the next five years will define the next economic era. Policymakers are fighting the last war. AI-related trade grew at 40 percent while governments debated the revival of 20th-century industrial policy. Second, trade agreements must be updated to address AI-specific realities. The current international trade architecture was built for goods that can be counted and services that can be defined. It has no adequate framework for AI models, training data, or the intellectual property embedded in the systems that now run much of the global economy. The WTO’s relevance is already questionable; without reform that addresses these new categories, it risks becoming entirely vestigial. Third, and most urgently, governments must grapple with the concentration problem. The AI-related trade is not distributed evenly. It flows between geopolitically aligned economies, a McKinsey finding with profound strategic implications. The world is sorting itself into AI-haves and AI-have-nots along lines that map closely onto existing geopolitical fault lines. Countries that fall on the wrong side of that divide do not merely lose economic opportunity; they lose the capacity to independently develop, deploy, and regulate the technologies that will define governance itself. The AI trade boom was not predicted by the officials who designed tariff policy, by the legislators who debated trade frameworks, or by most of the analysts who modeled global economic scenarios. That is not a minor oversight. It reflects a structural failure in how governments think about the economy, one that prizes the familiar and the measurable over the emergent and transformative. The 40 percent growth rate will not last forever. Technological waves eventually normalize. But the underlying infrastructure, which is building the datacenters, the chip supply chains, the talent concentrations, the platform dependencies, is being laid right now. Governments that treat this moment as a secondary concern do so at their permanent strategic peril. The AI Trade Boom Nobody Predicted was originally published in Artificial Intelligence in Plain English on Medium, where people are continuing the conversation by highlighting and responding to this story.