Short answer: Explosive fill is still a binding input for drones and shells, and the US leans mainly on one producer for HMX and RDX, the high explosives used in many US munitions. Holston Army Ammunition Plant in Tennessee is the only US producer, though West Point's Modern War Institute (MWI) reports that the Army qualified a second source it hasn't named (MWI). Regulation is one factor among several. MWI points further upstream, at precursor chemicals, as the tougher constraint. And a new Houston-based startup, Odin Energetics, is pitching shipping-container-sized plants that make explosives on site. Its numbers are all company claims so far.
Who makes RDX and HMX in the US?
- In 1980, five civilian RDX plants had about 119 million lb a year of capacity. By 2006, one site was left, Holston, making under 7 million lb (MWI).
- A late-1980s RDX line at Louisiana's army ammunition plant "never became a lasting second source" (MWI).
- An Army spokeswoman said Holston has been the only producer "since widespread closures of government-owned plants in the 1990s" (DroneXL, summarizing the Wall Street Journal's Oct 27, 2026 story "America Needs Explosives for Its Drones. This Startup Is Making Them").
- After a 2016–2020 demand surge, MWI says, the Army had to "qualify a second source it has not publicly named." So "the only US producer" needs that caveat.
Why the upgrade is slow: still an open question
Holston's modernization has slipped, by reported figures that are secondhand. The only official reason so far: the timeline "shifted to ensure safety and technical standards were being met" (Army spokeswoman, via DroneXL). Whether permitting, contracting, technical problems or funding drove it isn't established. We're not going to guess.
Where regulation does show up
Three ways, as one factor among several:
- Precursor controls. Acetic anhydride is internationally controlled as a heroin precursor. MWI calls it "the hardest link in the chain."
- Safety siting. The Army's own plan aims to "minimize the concentration of energetic materials in any one place" (USAASC).
- Local opposition. Odin pitches on-site production as a way around "local opposition to bomb factories" (DroneXL).
Decades of consolidation and uneven demand matter at least as much (our view).
The new entrant: containerized explosives plants
Odin Energetics, a joint venture of Solugen and Kraken Kinetics, pitches "Nitroforge" units that make nitramine explosives on site (Soldier Systems Daily). Odin claims about six months to build, output "close to a tenth" of Holston's, and $115M committed, per the WSJ as summarized by DroneXL. These are company claims. Solugen's own page still describes Nitroforge as "demonstration scale" (Solugen). Solugen's Terraforge precursor plant could show that precursors can be addressed, but what it will produce hasn't been confirmed (Solugen).
Do nested drones need more energy per gram?
Our working thesis is that nested drones (drone motherships that carry other drones) need more destructive power per unit of mass. That's untested, not proven. Odin says its formulation is "30% more powerful," without stating the baseline: it names no comparison explosive and no basis (per mass, per volume or per effect). DroneXL notes the figure "is the company's own." The best-known high-energy option is CL-20, invented by the Navy at China Lake, California, and "considered more powerful than HMX and RDX." Northrop Grumman makes it in Utah and believes it's the only company doing so at scale (Inside Defense). The FY2024 NDAA directed a CL-20 pilot on three weapons platforms; its status is unknown. We haven't found a verified CL-20-versus-HMX figure, so we're not giving one.
Two cautions, both our reasoning rather than findings: in a small drone, the explosive is only one share of the mass, so a better fill helps only that share. And nesting may be about surveillance or relaying signals, which doesn't depend on explosives at all. Whether nesting has been shown in combat is also open.
What it means for you
- Senior defense leaders and acquisition: the fill is a concentrated point of failure for drone and munitions plans. The Pentagon's Drone Dominance Program, with 60,000 attack drones (DroneXL, citing the WSJ), depends on this supply chain.
- Government-backed funds: a second, distributed source would lower that concentration risk. The test is qualification of new lots and formulations, not just a plant.
- Founders and engineers: if containerized units work, demand shifts to site-qualified, automated chemistry, which is a robotics and controls problem (our view). LA-area ties:
- Mach Industries' Exquadrum, in Victorville (Mach).
- Castelion, headquartered in Torrance, which builds its rocket motors in New Mexico.
- Anduril Industries, headquartered in Costa Mesa, which makes solid rocket motors in Mississippi.
- Kraken Kinetics' TERMINUS payload on the Archer drone from Torrance-based Neros Technologies (Kraken Kinetics).
- Investors: treat unbaselined performance claims like "30% more powerful" as a diligence filter.
Open questions
- Why did Holston slip, and who is the unnamed second source?
- How much domestic acetic anhydride capacity exists?
- Is "30% more powerful" per mass, volume or effect, and against which fill?
The questions we're asking: what would it take, in time, permits and money, to stand up a second RDX/HMX plant? And can small, automated, distributed units be qualified fast enough to matter?
Your turn: if you work in energetics, permitting or munitions acquisition, reply with what you're seeing. If you've seen a baselined figure for "30% more powerful," send it our way. Follow LA Robotics Network for updates.

