Most of the briefings in this series describe a company trapped by the labor shortage. This one describes a company trying to design its way out. Anduril's thesis is that the skilled-trades gap is partly a choice: the product of building weapons only a scarce, specialized workforce can assemble. Change the design, the argument goes, and you change who can build it. Arsenal-1 is that argument poured into concrete. It is the most credible corporate answer to the labor shortage anyone has tried. And it is about to meet the hardest labor market it could have chosen.
What is Arsenal-1?
Anduril announced Arsenal-1 in January 2025 on roughly 500 acres near Rickenbacker International Airport in Pickaway County, just south of Columbus. The plan is to reach about 5 million square feet across seven buildings, which would rank the site among the largest factories in the country. Anduril committed between $900 million and $1 billion and more than 4,000 jobs, and Ohio called it the largest single job-creation project in state history.
The factory builds autonomous systems: the Fury collaborative combat aircraft, the Barracuda cruise missile, and the Roadrunner interceptor, with a classified program setting up alongside them. Fury production began in early 2026, on schedule or ahead of it. Anduril brands the approach "hyperscale," a factory meant to mass-produce affordable autonomous weapons rather than a handful of exquisite ones.
| Arsenal-1 | Figure |
|---|---|
| Location | Pickaway County, Ohio (near Rickenbacker, south of Columbus) |
| Size at full build-out | ~5 million sq ft, 7 buildings, 500 acres |
| Investment | ~$900 million to $1 billion |
| Jobs committed | 4,008 direct, by 2035 |
| Builds | Fury (CCA), Barracuda missile, Roadrunner interceptor |
| Announced | January 2025 · production began early 2026 |
Investment and job figures are Anduril's committed targets under its Ohio incentive agreement, not realized totals. Sources: JobsOhio; Manufacturing Dive; Air & Space Forces Magazine; Breaking Defense.
How many workers does it need, and by when?
The incentive agreement commits Anduril to 4,008 direct jobs by the end of 2035, though the company has spoken about getting there faster. That distinction matters. Today the ramp is early: Fury production started with a small team, and Anduril aimed for about 250 workers on site by the end of 2026. The 4,000-plus number is a ten-year projection, not a payroll. The staffing question is not whether Anduril can find 250 people. It is whether central Ohio can supply a few thousand skilled production workers over a decade while everyone else in the region hires the same people.
Anduril's bet: design the labor problem away
Here is where Anduril departs from every other company in this series. Its leadership argues that traditional defense manufacturing built itself a labor trap. "The reality is that we have designed these luxury good items in defense," CEO Brian Schimpf has said. "These are artisanal, handmade products." Handmade products need artisans, and artisans are exactly the scarce, multi-year-trained workforce the industrial base cannot find.
So Anduril designs in the other direction. It builds with commercial off-the-shelf parts, roughly 70 percent of the Barracuda missile, and with commodity tooling. Schimpf says the Barracuda needs "something like 95% fewer tools to actually assemble." The company says it can pull workers from automotive, consumer electronics, and commercial aerospace instead of a narrow defense-trained pool, because it chose materials and processes "the broadest workforce possible" can handle. The Barracuda, by Anduril's account, takes about 30 hours and ten or fewer common hand tools to put together. Lower the skill required to build the thing, and you widen the pool of people who can build it.
It is a good idea, and the right instinct. Design for manufacturability is how Toyota and SpaceX out-produced their rivals. But be precise about what it solves, and what it does not. Designing for a broader workforce lowers the height of the skill barrier. It does not add anyone to the labor pool on the other side of it. Anduril still needs thousands of people who can show up, run a line, hold a tolerance, and pass a background check. The design makes each hire easier to train. It does not make the hires appear.
Why Ohio is the hardest place to test that bet
Anduril chose a region already at war over skilled labor. Within an hour's drive, three of the largest industrial projects in the country are hiring at once.
| Competing employer | Investment | Draw on the same labor pool |
|---|---|---|
| Intel (New Albany) | ~$28 billion | ~3,000 direct jobs plus ~7,000 construction trades |
| Honda / LG (Jeffersonville) | ~$3.5 billion | ~2,200 jobs at a battery plant |
| Data centers (Google, AWS, Meta, Microsoft) | Billions | Heavy draw on construction electricians and trades |
Intel's Ohio fabs were delayed to 2030-2031 as of early 2025, which eases near-term competition for permanent fab technicians but not for construction trades. Sources: Manufacturing Dive; JobsOhio; LG Energy Solution; Data Center Frontier.
Now set that demand against the supply. The Columbus metro has roughly 76,000 manufacturing workers, and that number has barely moved in years. Manufacturing is only about 6 percent of the area's jobs, a small base to draw a few thousand skilled hires from. Regional unemployment ran near 4 percent through 2025 and tightened further into 2026. When a market is this tight, adding a 4,000-job factory does not fill the jobs so much as bid up the price of the workers who already have them. Anduril's posted production wages, starting around $24 to $31 an hour, land in a market where Intel's construction trades can clear six figures with overtime.
The training pipeline is expanding, but it was built for the neighbors. Columbus State leads an eleven-college consortium aimed at preparing thousands of semiconductor technicians for Intel, and the building trades have grown their apprenticeships to feed construction. Both are welcome. Neither was designed for defense production, and both still run behind the region's combined demand. Anduril's answer is that its roles overlap with automotive and electronics work, which is true, and which is also the labor every other employer in the region chases.
The constraint the others do not have
Intel, Honda, and the data centers can hire almost anyone qualified. Anduril cannot. Arsenal-1 builds export-controlled defense hardware under ITAR, so most production roles require US-person status, a US citizen or lawful permanent resident, and many require the ability to hold a security clearance. That single requirement carves a large share of the labor pool out of Anduril's reach in a market already short. It is the same wedge that binds the shipyards and the fabs supplying defense: the clearable workforce is a subset of a shortage, and a subset is smaller than the whole.
The national backdrop
Arsenal-1 is one node in a much larger buildout. The Manufacturing Institute projects a shortfall approaching 1.9 million manufacturing workers by 2033. Shipbuilding alone needs roughly 174,000 new workers over the decade. Ohio has projected its own shortage of more than half a million skilled workers and stood up a workforce initiative to match. Anduril is not a startup betting against these numbers from the outside. It is now a roughly $61 billion company that has grown past 8,000 employees and doubled revenue in a year, which means its appetite for skilled labor is compounding, not leveling off. The company that most wants to rebuild the arsenal of democracy is discovering, in Pickaway County, that the binding input is not capital or contracts or even factory space. It is people who can build.
The bottom line
Anduril's design-for-manufacturability bet is the smartest response to the skilled-labor shortage any large employer has made. It should lower training time, widen the candidate pool, and let the company hire faster than a traditional prime ever could. All of that is real. But design is a multiplier on a workforce that has to exist first. Four thousand jobs, dropped into a flat base of 76,000, against Intel and Honda and the data centers, under an ITAR filter the others do not carry, is not a problem you finish solving in the engineering department. Somebody still has to build the people. That is the work we do at Project Arklight, and it is the constraint Arsenal-1 will spend the next decade proving cannot be skipped.