The U.S. Industrial Fork in the Road
The United States is restructuring its energy infrastructure, its manufacturing geography, and its workforce composition at the same time, against a backdrop of demographic aging. Three forces are arriving together, and the next decade will determine whether we get broad industrial expansion or something narrower and more concentrated.
I wrote a full paper on this earlier in the year. What follows is the argument in short form, with the practitioner's version of what it means if you run sourcing or operations at a mid-sized manufacturer.
AI is a multiplier, and the gap is widening
Artificial intelligence is the most significant productivity engine since the commercial internet. In supply chain work specifically, it has already changed the economics of tasks that used to consume days: HTS classification research, contract review, landed-cost modeling, freight routing analysis.
The gains are real and they are large. But they are not evenly distributed, and that is the part worth paying attention to. Firms that have built these tools into how they actually work are pulling away from firms still running the same analytical processes they ran in 2020. The competitive distance between the two groups is growing faster than most people realize, because the advantage compounds — every hour saved gets reinvested into work that generates more advantage.
This is a leverage tool, not a replacement for judgment. The tools are fast. Knowing what to trust is still the job.
Energy is the binding constraint
In an AI-driven economy, electricity is the physical reality of productivity. Data centers and high-automation factories need reliable baseload power, and a single large AI facility can draw as much as a mid-sized city.
For anyone planning a facility expansion or evaluating a reshoring decision, this stops being a policy abstraction. Grid capacity and permitting speed are now critical-path items — the regions that solve transmission and permitting first will capture the next wave of industrial investment, and the regions that don't will watch projects go elsewhere regardless of what else they offer.
The barbell economy
A bifurcated pattern is forming. Large corporations are deploying AI to consolidate and reduce headcount in administrative and analytical functions. At the other end, small technology-enabled operators are multiplying their output without adding staff, and competing for work that used to require a firm ten times their size.
The middle is where the pressure lands. Mid-sized manufacturers slow to adapt face compression from both directions at once — undercut on cost by the scaled players, and outmaneuvered on speed by the small ones.
Tariffs stopped being a fixed cost
The 2025–2026 tariff environment is the most significant restructuring of U.S. trade policy in close to a century. Section 301 and IEEPA-based measures have changed sourcing math across the board.
The specific problem worth naming is duty stacking — the accumulation of multiple tariff layers on a single product. Managing it requires granular work across three fronts: tracking how standard MFN, Section 301 and Section 232 rates combine on a given classification; systematically mapping alternative origins; and modeling the landed-cost impact before the input cost shift hits the margin.
Tariff complexity is an opportunity. Winning companies actively manage duties through accurate classification and first-sale valuation.
This is the shift I'd emphasize most: customs and trade compliance has moved from a cost center to a value-creation function. Companies still treating classification as a clerical task are leaving real money on the table.
Nearshoring is further along than the commentary suggests
Automation is rewriting the logic that drove decades of offshoring. As robotics reduce the labor share of production cost, resilience and proximity start to outweigh the wage differential.
A North American pattern is forming around comparative strengths — the U.S. for capital-intensive, high-automation and IP-sensitive production; Mexico for industrial assembly and mid-complexity manufacturing; Canada for energy, resource processing, and integrated industrial supply.
In my work with clients restructuring supply chains away from China, Mexico has become the dominant alternative. The questions have shifted from "why consider nearshoring?" to "how do we qualify suppliers and manage compliance?"
That shift in question is the whole story. The decision has largely been made; the work now is execution — supplier qualification, USMCA compliance, and building the operational infrastructure to support it.
Two paths
The forces above point toward genuinely different outcomes, and which one we get depends on choices that haven't been made yet.
In the better case, energy supply expands decisively, manufacturing regionalizes into a coherent North American ecosystem with tariff policy providing the incentive, and AI productivity spreads across the full economy — letting smaller firms serve more clients at higher quality.
In the adverse case, permitting bottlenecks delay grid upgrades, China retains dominant manufacturing share while tariff policy creates cost drag without successfully incentivizing domestic substitution, and AI gains accrue mainly to the largest technology companies.
The levers that decide it are identifiable: permitting reform for energy transmission, pairing tariffs with actual industrial capacity investment, vocational training and targeted skilled immigration, and housing supply in the regions where the work is.
What this means if you run sourcing
Three things I'd act on regardless of which path we end up on.
Treat resilience as a strategic priority rather than an insurance policy. The firms that rebuilt for robustness after 2021 are the ones absorbing the current volatility without drama.
Reframe trade compliance as value creation. Classification accuracy, origin strategy and valuation method are margin decisions, not paperwork.
Build AI into procurement and sourcing work now. Not to replace the expertise — to raise the output-per-head ratio enough that you stay competitive with firms on both ends of the barbell.
The fork is real. Adaptation is required at the policy, organizational and individual level, and the window to choose a trajectory is open now.
Get the full paper
The full white paper goes considerably deeper on each of these — demographics and the worker-to-retiree ratio, global industrial competition and China's robotics position, the complete Path A / Path B comparison, and the five policy levers with sourcing notes.
Sources: U.S. Social Security Administration worker-to-retiree projections; Bureau of Labor Statistics manufacturing employment and output data; International Federation of Robotics global installation statistics; U.S. Energy Information Administration electricity demand data; USTR Section 301 actions and the Harmonized Tariff Schedule.
Mike Dates is the founder of Supply Chain Works LLC, an independent consulting practice focused on global sourcing, tariff analysis, HTS classification and import/export compliance.
Comments