Mach Industries Bets on Solid Rocket Motor Mass Production

Mach Industries CEO Ethan Thornton sat down for a nearly two-hour conversation on the Relentless podcast to explain, among other things, how a Los Angeles defense startup plans to build one of the largest solid rocket motor factories in the United States. The conversation covered ground from hypergolic propellants to block redesign philosophy to the unit economics of asymmetric warfare. Thornton is either building one of the most important new defense industrial companies in America or spreading himself so thin that something breaks. Probably both, depending on the quarter.

The rocket motor ambition is the most concrete anchor in an otherwise sprawling portfolio. Mach is retrofitting roughly 200 bunkers near the old George Air Force Base, about 90 miles east of Los Angeles, to store propellant precursor. The current factory runs 117,000 square feet; Thornton says a second facility, roughly twice that size and dedicated specifically to energetics, was targeted to open within four months of the recording. The eventual production target is hundreds of thousands of solid rocket motors per year, a figure that would represent a meaningful share of total U.S. output from a company that did not exist in this form a few years ago.

The chemistry problem is not trivial. Thornton describes solid rocket motor manufacturing the way a pastry chef might describe bread: "jet engines are plumbing, solid rocket motors are baking." You mix a dough-like propellant, pour it into a casing, cure it at precise temperatures, and hope it neither fails to ignite nor ignites at the wrong moment. The process is simultaneously low-tech in its analog steps and extremely unforgiving in its failure modes. "By the time you light them, you don't have really any control," Thornton says. "It either works or it doesn't work. And when it doesn't work, it is usually quite kinetic."

To get into the business without waiting five years for the required federal licensing, Mach acquired a company called Squadron, a two-decade-old design lab founded by former Air Force solid rocket motor engineers who had spent years winning contracts from the Air Force, NASA, DARPA, and SOCOM. It was one of the few acquisitions Thornton says he does not regret, precisely because there was no viable path around it. The licensing timeline alone would have foreclosed the whole plan.

The supply chain situation is arguably more fragile than the manufacturing chemistry. Thornton says the precursor for most U.S. solid rocket motor production currently comes from a single domestic facility that is booked out roughly three years. That single-point-of-failure keeps procurement officials awake and makes Mach's ambition to bring precursor production on-site, potentially partnering with suppliers to co-locate, look like strategic necessity rather than vertical integration for its own sake.

TechCrunch has noted that Thornton is pursuing an unusually broad product portfolio simultaneously: drones, missiles, kinetic interceptors, high-altitude balloons, solid rocket motors, and warheads, alongside programs not yet publicly announced. Thornton acknowledges the distraction critique directly. His defense is that in defense specifically, a multi-product company built from day one on functional (rather than product-siloed) engineering teams actually scales more cheaply than a single-product company that later tries to diversify. Whether that holds at the manufacturing scale he is targeting is an open empirical question.

The core strategic logic underneath all of it is a specific theory about how the United States competes with China industrially. China outbuilds the U.S. on commercial shipping tonnage by roughly 232 to 1. Thornton's answer is to not compete on that axis at all. A single shipping container can deploy a meaningful number of Pike cruise missiles. Each Pike can, in his framing, neutralize a small ship, and three can handle a larger one. The math: if Mach can produce tens of thousands of Pikes per month at a unit cost far below the value of what they hit, the U.S. has manufactured asymmetry without closing the shipbuilding gap. "If each pike can take out a small ship and if three pikes can take out a big ship, suddenly I have created asymmetry against China's advantage of building more ships than us."

The honest constraint on this logic is iteration speed. Thornton expects China to copy Pike and engineer a countermeasure, probably within months to two years. Mach is already designing Dart, its own interceptor, explicitly modeled on what a Pike-like system looks like when it has been reverse-engineered by an adversary. The goal is to have the counter-to-the-countermeasure ready before the countermeasure is deployed. This requires a product development cycle measured in months, not years. Mach claims its current average, from whiteboard to first flight, is around one year, down from two. Thornton wants to compress that to two months.

The jet engine development story is the clearest illustration of how he tries to do that. Facing quality and supply problems from external vendors and a Western production rate of roughly 300 small jet engines per month against customer demand measured in the tens of thousands, Thornton ran two parallel engine development programs simultaneously. One team, led by engineer Jeremy Clyde, pursued a high-performance, high-risk design with a bill of materials of 30 to 50 major components, optimized for rate manufacturing. A second two-person team, working on a $200,000 budget, built something that looked closer to existing market options. The low-risk engine fired in seven months. The high-risk engine fired in eight. Both were completed within roughly the one-year window that supply chain constraints demanded. The low-risk design won on the combination of cost, performance, and timeline.

On the financing structure, Thornton pushes hard toward operating expenditure over capital expenditure, for reasons specific to the defense procurement cycle. A company cannot book production revenue until the government issues a production contract. But the government often will not issue a production contract until the company demonstrates production capacity. The company therefore must pre-invest in capital before it has the revenue to justify it. Keeping capital expenditure light and operational expenditure variable lets Mach defer some costs until contract execution, and it preserves surge capacity: the ability to scale rapidly when an order arrives rather than being locked into whatever factory footprint existed at contract signing. The Wall Street Journal has reported on a $50 million contract in Mach's portfolio tied to hypersonic development, which gives some sense of the revenue scale the company is operating at while it builds toward production.

What Thornton describes as the "valley of death" in defense, the stretch between a product at 90% readiness and the government actually ordering production, is where he says most companies accumulate overhead and bloat, adding engineering complexity to justify continued development contracts rather than pushing toward rate. His stated solution is to always have products at multiple stages simultaneously, so that engineers who specialize in early-phase development do not leave when one program matures, and so that the company never becomes purely a development shop waiting on a procurement decision.

The questions Thornton does not fully resolve in the conversation are the ones worth tracking. Can a factory optimized for flexibility actually achieve the unit economics of one optimized for rate? The glide airframe example he cites, a redesign that dropped cost from several thousand dollars per unit to $7 per airframe and cut operator hours from 50 to near zero, is striking. But that number covers the airframe only; add warheads, avionics, and propulsion, and the cost picture changes substantially. The precursor supply problem is structural to the whole industry, and finding two or three partners willing to co-locate production is a diplomatic and logistical effort that has nothing to do with engineering talent. And the government still controls when production contracts get issued, which means the most important variable in Mach's revenue model is not Thornton's iteration speed but a procurement bureaucracy operating on its own timeline.

For now, Mach is retrofitting 200 bunkers in the Mojave desert, running flight tests of Pike at roughly one per day during active campaigns, and building jet engines in San Luis Obispo because the LA Air Quality Management District would not permit engine testing in Los Angeles. The company that TechCrunch describes as trying to do everything all at once is, at minimum, doing a lot of things at once. Whether that is visionary industrial architecture or a very expensive way to find out which bets were wrong is the question the next 18 months will start to answer.

By Elena Vasquez-Moreno