Zoffer: Use Data-Center Demand to Rebuild US Supply Chains

Every big AI-infrastructure story right now gets framed either as evidence of American technological confidence or as a burden on grids, water, and local politics. A new Financial Times op-ed by Josh Zoffer tries to change that frame, arguing that data centers are actually one of the more valuable industrial-policy anchors the US has stumbled into in a generation.

Zoffer's core argument, flagged on Techmeme, is that data centers offer the US a chance to get ahead in the next key technologies and to build domestic supply chains based on demand rather than subsidies and tariffs, and that America must not repeat the mistakes that handed China rare-earth dominance. He is writing from a policy background that gives the framing some weight: from 2023 to 2024 he served as Special Assistant to the President for Economic Policy at the White House National Economic Council, working on energy security, trade policy and supply chain issues, and he now leads research at CIV, an investment firm that builds and backs technology companies in critical industries, and is a Senior Non-Resident Fellow at Columbia's Center on Global Energy Policy.

The interesting move is the reframe. The default DC playbook for a strategic industry has been to write a subsidy program, layer on tariffs, and hope domestic producers scale before the political oxygen runs out. Rare earths are the canonical example of what happens when that playbook fires too late. Zoffer's counter is that AI data centers are already generating the kind of demand pull that rare-earths policy has spent years trying to manufacture, and that if the US lets that demand flow through to domestic suppliers, it gets industrial capacity without a fresh subsidy fight.

The honest caveat is that a lot rides on words the argument uses lightly. Demand does not automatically translate into domestic production if hyperscalers can source cheaper and faster abroad, and the material visible outside the FT paywall doesn't spell out which policy levers would actually route procurement toward US suppliers, or which specific downstream industries Zoffer thinks data-center demand would anchor. The mechanism is left as assertion more than blueprint.

Still, the framing is worth taking seriously if you work anywhere near energy or industrial policy. Read this way, the AI capex cycle is the industrial-policy opportunity, not an externality to be managed around, and the people best-positioned if the argument lands are the US-based makers of the physical inputs data centers consume, along with the investors, Zoffer's own CIV among them, already lined up on that thesis.