The Arsenal of the Midwest
The Hidden Bottleneck of National Security
My father spent twenty-eight years on the line at Pontiac Assembly. He came home most nights smelling of paint and metal in a neighborhood where nearly every other father did the same. The street ran on General Motors time. Shift changes set the rhythm. First, second, or third shifts. What I understood growing up, without anyone having to explain it, was that the work was not separate from the life.
Then the plants closed. Not all at once, which might have at least clarified things, but plant by plant, supplier by supplier, across the nineties and into the aughts, until the industrial ecosystem that had sustained Flint for three generations was gone. What I remember most is not the anger, though there was plenty, but the particular eeriness of an abandoned factory. A building still standing, windows intact, parking lot empty on a Tuesday morning.
I left for college and then the Army, spent years in the infantry learning that logistics is not a support function. It is the war itself. A unit fights at the level its supply chain allows. I carried that lesson home. Reading defense procurement reports as an adult, I keep arriving at a conclusion that feels less like analysis than memory: what happened to Flint is not a social tragedy separate from our national security. The two things I know best in the world have turned out to be the same problem.
Consider the timeline.
In 2021, Admiral Phil Davidson testified before the Senate Armed Services Committee that China was building the capability to take Taiwan by force, and that the threat was “manifest during this decade, in fact in the next six years.” CIA Director Bill Burns later confirmed that Xi Jinping had instructed the People’s Liberation Army to be ready to invade by 2027. The date is a planning horizon, but planning horizons have a way of becoming operational realities. This one arrives in less than one year.
The reason Taiwan matters to the defense manufacturing question is what it produces. TSMC alone controls roughly 64 percent of global semiconductor foundry revenue. Taiwan accounts for over 90 percent of the world’s most advanced logic chip capacity. Every modern weapons system is, at its core, a computing device. The F-35’s avionics, the Navy’s fire control systems, the guidance packages on precision munitions: all of them run on chips that, in their most advanced forms, only one island on earth can currently make.
A blockade or invasion would sever the nervous system of the modern military. The United States, despite the CHIPS Act and billions in domestic investment, remains years from having the fabrication capacity to compensate. The bottleneck is the trades. Semiconductor fabrication plants are built and maintained by pipefitters, millwrights, sheet metal workers, and precision machinists; the same blue-collar workforce that built the auto plants of Flint and has been told for thirty years that its skills belong to a dying economy. We have plenty of engineers who can design the fabs. We do not have enough tradespeople who can build and run them.
China understands this and has been methodical in using it. F-35 production was halted for months after Chinese-made magnets were discovered in the aircraft’s supply chain. The dependency was by design. Not China’s design. Ours. It was the residue of thirty years of decisions made by people who were somewhere else, optimizing for cost and convenience while the manufacturing base that might have offered an alternative was going dark in the Rust Belt.
Ukraine showed the first cracks. Since February 2022, Ukrainian forces burned through artillery shells at a rate no Western planner had seriously modeled. At peak intensity, Ukrainian gunners fired upward of 7,000 rounds per day. When the war began, the United States Army was producing roughly 14,000 shells a month.
Fourteen thousand. Less than two days of Ukrainian battlefield consumption from the entire output of American industrial capacity.
The Pentagon has spent nearly $5 billion trying to correct this. New factories have opened in Texas, Arkansas, and Kentucky. The monthly target, set in early 2024, was 100,000 complete rounds by October 2025. As of mid-2025, production stands at roughly 40,000. The Army now says early 2026, and reaching that number requires 66,000 tons of explosives per month, much of it previously imported because the United States stopped manufacturing TNT domestically in the 1980s.
Then came Iran. The strikes expended stockpiles the industrial base had not yet rebuilt from Ukraine. According to Payne Institute estimates, the U.S. is now roughly six weeks from exhausting its Tomahawk inventory at current replenishment rates, two weeks on Patriot PAC-3 interceptors, and thirty months from replacing its GBU-57 bunker busters, the one weapon specifically designed to destroy hardened underground facilities.
The Davidson Window opens in 2027. The replenishment chart extends to 2028.
Those two facts, sitting next to each other, are the whole argument.
A generation of betting on the knowledge economy hollowed out our ability to build things. We traded the trades for degrees, the factory floor for the open office, the machinist for the consultant. It made sense on a spreadsheet. The defense implications were someone else’s problem, until they weren’t.
What rarely gets said plainly is that the bottleneck is not the primes. It is the two hundred thousand suppliers beneath them. The tier-two and tier-three operations making forgings and castings, propellant charges and fuzes, the unglamorous parts that make the weapons work. Those suppliers are in places that look like Flint. And Flint has been losing them for thirty years.
The Rust Belt is a graveyard of tier-two and tier-three suppliers running at subscale, often owned by founders without successors. These shops sit on specialized equipment and decades of tribal knowledge that the defense industrial base urgently needs, yet they are invisible to the modern economy. They never made it into a venture pitch deck because they are boring. But in large-scale combat operations, boring is what scales.
The CHIPS Act is part of the plan, but we need a succession plan for the American machinist. These small-town shops are the latent Arsenal of the Midwest, currently held together by a generation of 70-year-old owners who are one retirement away from taking our surge capacity with them. Reclaiming this ground requires a tactical re-occupation of the shop floor by a generation that understands both the spreadsheet and the supply drop. It requires patient capital and a reenergized focus on the last tactical mile of production. We must look at a machine shop in a mid-sized Michigan city and see not a relic of the 20th century, but the strategic high ground of the 21st.
The Davidson Window opens next year while we are still chasing the shell production target. The TSMC fabs in Arizona remain a multi-year construction project. The need is undeniable, the timeline is short, and the assets we need are hiding in plain sight, tucked into the same valleys where my neighborhood once set the heartbeat of American manufacturing.
I spent years in uniform learning what happens when the supply chain isn’t there when you need it. My father spent twenty-eight years at Pontiac Assembly proving it doesn’t have to be that way.
We knew how to do this once. The question is whether we remember in time.