The last time anyone counted, the United States had roughly eight days worth of the weapons it would need in a grinding fight against China. Not eight months. Not eight years. Eight days.
That number comes from Palantir CTO Shyam Sankar, the company's thirteenth employee when he joined in 2006, and it is the number that explains why Anduril, the defense startup co-founded by Trae Stephens, just opened a one-billion-dollar factory south of Columbus, Ohio. According to Reuters, the Arsenal-1 manufacturing campus covers five million square feet and is expected to employ more than four thousand people over the next decade. It builds drones designed to be cheaper than the missiles aimed at them. And it represents an answer to a question the US defense establishment has been avoiding since Ukraine burned through ten years of American weapons production in ten weeks of fighting.
Ukraine showed what modern attritional warfare actually costs in hardware. "Our adversaries are not coming at us with ten-plus million dollar fighter planes," Stephens told Fox Business. "They are coming at us with very, very low-cost munitions." When your enemy can spend fifty thousand dollars on a drone that destroys a tank, the math of deterrence stops working.
Anduril's answer is the YFQ-44A Fury, a collaborative combat aircraft that flies alongside crewed jets. Per Defense One, Anduril started armed flight testing of the system in February 2026 and is one of three companies competing for the Air Force's Collaborative Combat Aircraft program, alongside General Atomics and Northrop Grumman. The first product off the line at Arsenal-1 will be the Fury.
The production line can build up to fifty aircraft per shift, according to Jason Levin, head of production at Anduril. Per Inside Unmanned Systems, at three shifts per day that works out to a theoretical maximum of one hundred and fifty aircraft per year. It is a number Anduril's marketing materials emphasize. It is also a line rate, not an order volume. The Air Force has not committed to buying one hundred and fifty of anything. The CCA competition is still running. What Anduril can build and what the Pentagon will actually pay for are two different questions.
Stephens has made the industrial base argument in multiple forums. His framing is blunt: Congress has done nothing on immigration despite seventy to eighty percent public support for reform, he told Reuters. The implication is that waiting for the legislative branch to fix the weapons stockpile is not a strategy. Anduril building the factory is the path.
There is something to that argument. Sankar made the eight-day figure and the WWII precision statistic on Fox News: in World War II, only six percent of bombs dropped hit their targets. Modern precision-guided munitions changed that calculus, but they are expensive and slow to produce. Ukraine demonstrated that attritional conflict consumes industrial output at a rate no peacetime supply chain was designed to sustain. His broader point was that US weapons stockpiles had been drawn down without being replenished at the rate peer competition requires.
The US Air Force, during its confrontation with Iran, conducted two thousand strikes in forty-eight hours using AI-enabled systems. Per Fox News, planning happened in a fraction of the time prior conflicts required and strike volume per day was more than double what similar-scale operations produced a generation earlier. This is the future Sankar and Stephens are selling: not a few exquisite platforms, but swarms of capable, cheap, replaceable ones.
Whether Anduril can deliver that future is another question. The company has a track record of bold timelines that compress under pressure. Arsenal-1 was originally scheduled to begin production in July 2026. Per Defense One, production is imminent as of March 2026, which suggests either genuine acceleration or retroactive calendar adjustment. The four thousand jobs over the next decade are a mix of current hires and future headcount, the kind of window that lets a company claim credit for employment that does not yet exist.
The people working the line at Arsenal-1 are part of this story too. Four thousand manufacturing jobs in southern Ohio is a deliberate choice. Anduril is not subtle about the political dimension. Stephens has argued that legislative failures on immigration and industrial policy are why the private sector had to act. The factory is the argument.
What happens next depends on what the Air Force decides. The CCA competition will determine which companies get the first major production contracts for collaborative combat aircraft. Arsenal-1 is designed to scale if those contracts materialize. If they do not, the factory becomes an expensive monument to a theory of warfare that has not yet been proven at the tempo Anduril is promising. The drones are real. The question is whether anyone in Washington will buy enough of them, fast enough, to matter.