Jensen Huang no longer describes data centers as warehouses for information. NVIDIA’s CEO calls them “AI factories” — industrial systems that convert electricity into intelligence. That language matters because it exposes a truth policymakers spent decades trying to ignore: the digital economy was never truly digital. It is an industrial system disguised as software.
And that is why the real contest between nations is no longer about tariffs or semiconductor restrictions. It is about infrastructure.
What I call the “Geopolitical Tax” is the hidden surcharge paid by nations that allowed their energy systems, transmission networks, and water infrastructure to stagnate while their digital ambitions accelerated. We are entering an era in which the greatest obstacle to economic expansion may not be a tariff or a sanction, but a transformer that takes four years to manufacture and another two years to connect.
To “work their magic,” AI systems require hyperscale data centers that consume staggering amounts of electricity and water. The International Energy Agency warned this spring that electricity demand from data centers surged 17% in 2025 alone and could double by 2030, while AI-focused facilities may triple their power consumption during the same period. Supply chains for transformers, turbines, grid equipment, and transmission infrastructure are becoming critical bottlenecks.
That logjam is no longer theoretical. Utilities from Virginia to Arizona are warning that data-center growth is outpacing local grid capacity. In some regions, projects face years-long delays simply waiting for interconnection approval. Meanwhile, residents increasingly fear rising electricity rates and mounting pressure on local water supplies.
The AI revolution, in other words, is colliding with the physical world.
“AI is now infrastructure, and this infrastructure, just like the internet, just like electricity, needs factories,” Huang said, as reported in a blog. “These factories are essentially what we build today. They’re not data centers of the past. These AI data centers, if you will, are improperly described. They are, in fact, AI factories. You apply energy to it, and it produces something incredibly valuable, and these things are called tokens.”
From Algorithms to Assets: The New Geopolitical Race
That language matters because it reframes AI not as a software sector but as heavy industry.
For decades, globalization rewarded nations that optimized software, logistics, and low-cost manufacturing. The next era may reward countries that can build substations, transmission corridors, and generating capacity faster than their rivals. Countries such as India are treating energy infrastructure as a strategic asset rather than a regulatory burden; it has recently surpassed the United States in annual solar installations.
The geopolitical implications are profound. In the 20th century, great powers exported oil, manufactured goods, and financial capital. In the 21st century, they export grid stability — the ability to grow and transform domestic economies.
Europe’s post-Ukraine energy shock further exposed how vulnerable advanced economies become when infrastructure strategies fracture. Germany’s industrial slowdown and soaring electricity costs demonstrated that technological sophistication alone cannot compensate for unstable energy systems.
This is where much of the Western conversation about AI remains detached from reality. Policymakers still speak as if artificial intelligence floats above physical constraints, existing somewhere in “the cloud.” But there is no cloud. There are server farms, cooling systems, transformers, substations, gas turbines, transmission lines, and increasingly contested water resources.
And those systems are becoming harder to build.
From the hills of West Virginia to rural Iowa, communities are discovering that hyperscale data centers consume enormous amounts of electricity while competing for land and water. Residents worry — often correctly — that they will bear the environmental and financial costs of infrastructure expansion while technology companies reap the profits. In some cases, utilities are already proposing billions in grid upgrades tied directly to anticipated data-center demand.
But the absence of a national strategy creates an even more dangerous outcome.
Without coherent federal planning, the issue is not whether AI infrastructure will be built. It already is. The question is whether it will be integrated into a broader national development strategy or left to a patchwork of private deals and local political fights.
Why AI Demands an Eisenhower Moment
This is where the historical parallel becomes unavoidable.
When President Eisenhower championed the Interstate Highway System in the 1950s, Americans understood that roads were not merely local amenities. They were instruments of military logistics, industrial expansion, and national cohesion. The interstate system succeeded because the country recognized that physical connectivity was inseparable from economic power.
Today’s equivalent is not asphalt. It is transmission infrastructure, transformer manufacturing, modular nuclear reactors, water-recycling systems, and continent-scale fiber networks. These are the roads of the AI economy.
That reality also demands a more mature energy debate. The transition to cleaner energy remains essential, but modern economies cannot operate on slogans or ideological purity. Natural gas, nuclear power, geothermal energy, battery storage, solar, and wind are no longer competing political identities. They are components of what increasingly looks like a national security architecture. Sam Altman recently conceded that growing AI use will require massive new energy expansion and accelerated deployment of nuclear, solar, and wind generation.
The nations that dominate the next era may not be those with the best algorithms. They will be the ones capable of building physical systems faster than bureaucracies can obstruct them or geopolitical rivals can replicate them.
Economic power is shifting back toward countries capable of executing large-scale infrastructure strategies amid technological upheaval.
In the emerging multipolar world, the ultimate reserve currency may not be the dollar. It may be the reliable electron.
And if the United States wants to remain the country others organize around — rather than the one they merely work around — it must stop treating infrastructure as an afterthought to innovation and start recognizing that the age of artificial intelligence is, above all else, an infrastructure race.
