Walter Kemmsies' China+Many framework: e-commerce at 24 percent of retail, and what that means for the next industrial site search

Twenty-four percent of American retail now moves through a screen instead of a storefront, on the numbers Walter Kemmsies works from, and his projection carries it toward 30 percent by 2030. Every point of that shift converts selling floor into logistics floor somewhere inland. That alone would make industrial demand structural rather than cyclical.

It is only half of his argument.

Kemmsies is Managing Director and Chief Strategist of JLL's U.S. Ports, Airports and Global Infrastructure group, and the resume behind the framework is forty years deep: chief economist at Moffatt & Nichol, strategy roles at JPMorgan in London and UBS in Zurich and London, testimony before the U.S. Senate on freight competitiveness, advisory work for the Department of Commerce. The phrase he built for the second engine is "China+Many," his shorthand for supply chains diversifying production beyond a single-country manufacturing base. When production fragments across many origins, two things happen to American dirt at once. The import map redraws itself across a wider set of ports and inland corridors, and some of the production lands on this continent as an actual factory, which is a demand source no distribution-center model ever had to price.

So the framework tells you the demand is durable and roughly where it will flow. What it cannot tell you is whether a specific site will keep its promises. For that, look at the two biggest China+Many bets in our industrial files, both Korean manufacturers, both on ground that never needed a rezoning, and both instructive in what went sideways anyway.

Two campuses, zero rezoning fights

Samsung selected Taylor, Texas on November 23, 2021 for an advanced logic fab, initial stated investment around $17 billion. The entitlement story is a stack of executed agreements rather than contested hearings: a Williamson County Chapter 381 agreement, City of Taylor abatements with an expanded Tax Increment Reinvestment Zone, Taylor ISD arrangements, all sequenced against the federal award that landed as preliminary CHIPS terms on April 15, 2024 and a final $4.745 billion direct funding award on December 20, 2024, with conditions tied to milestones, workforce, and U.S. supply-chain provisions. The "Samsung Highway" road improvements opened ahead of their original schedule. The fab itself slipped, with risk production targeting 2026 amid the same capex-cycle pressure the whole industry is carrying. The case file scores it 72/100: approval path clear, schedule and abatement politics the live risks.

SK is the sharper lesson. SK Innovation picked Commerce 85 Business Park in Jackson County, Georgia in March 2018 for its first U.S. EV battery plant, $1.67 billion and 2,000 projected jobs, on land pre-zoned industrial inside a master-planned park. No rezoning, no conditional use permit, no variance. Construction started about a year later, a second plant announced in May 2020 took the commitment to $2.6 billion, and Georgia's Quick Start program had custom training curricula built for SK's production lines before the plant opened. Then in 2021 the U.S. International Trade Commission issued a decision in LG Chem's trade-secrets case that threatened SK's ability to operate in the United States at all. Governor Brian Kemp publicly lobbied the Biden administration, LG and SK settled, and the plant survived. Mass production began in January 2022 for the Ford F-150 Lightning and Volkswagen ID.4, hiring passed 3,000 jobs in 2023, roughly two years ahead of the original commitment. Then U.S. EV demand growth slowed, and the file records 958 layoffs across 2024 and 2025, taking the workforce to roughly 1,600.

The zoning held. The employment promise did not.

What the framework prices, and what it leaves on your desk

Notice what never failed in either case. Not the entitlement, and not Kemmsies' demand thesis either; the fabs are rising in Taylor and the cells still ship to Ford and Volkswagen. Everything that actually bit lived in the layer between the macro call and the local approval. A federal trade tribunal nearly stranded a fully-approved $2.6 billion campus, and it took a governor escalating to the White House to unwind it. A demand-side wobble in one product category erased almost half a workforce without touching a single permit. A global capex cycle moved a fab's production date while its highway opened early.

That layer is where the next site search will be won. The China+Many demand engine does not select your parcel. Both of these companies went where the ground was already prepared, a pre-zoned master-planned park in Commerce, an incentive machine already assembled in Williamson County, and my read is that this was not incidental to the site selection but the substance of it. The fragmentation trade rewards jurisdictions that did their entitlement work before the announcement, because a manufacturer choosing among many countries has no patience for a two-year rezoning fight in any one of them.

My bet, stated so you can check it: the binding constraint on China+Many manufacturing siting through 2030 will be pre-entitled land with secured power and water, not tariffs and not demand, and the announcements will show it. Take the next ten U.S. plant announcements above $1 billion and count how many land on pre-positioned sites against how many take on a genuinely contested rezoning. If even three of the ten pick the contested path, I called this wrong. I do not think you will find three.

This analysis is a source-cited research summary drawn from public records and industry reporting, not legal advice. It can contain errors and should be verified independently before any investment decision.