


Loading the Elevenlabs Text to Speech AudioNative Player...
Justin Marston is the founder and CEO of Mithril Defense, a drone startup based in Texas that aims to provide security solutions for domestic law and order problems. Its flagship program includes drones that are armed with pepper sprays to neutralize school-shooting suspects, the first such project in the United States. It’s a new company and boasts a team that includes a Seal Team Six command chief. Marston sat down with The American Conservative to answer some questions on why drones are the future, where we stand in a competition with China, and what the main challenges facing drone startups in the U.S. are.
What are the challenges facing the U.S. when it comes to drone swarms?
China totally dominates the drone industry. DJI makes over 70 percent of the drones that are purchased, and likely has components in half of the rest. You really have to purposefully design to be non-Chinese, otherwise some of the chips in the drone will have come from China. This is true in Ukraine and Russia too—the vast majority of the drones made in both countries have some if not the majority of their components coming from China, even if the drone itself is assembled in the local country. It's difficult to understate how much China is the center of gravity for the drone industry.
All the drones currently manufactured in the U.S. are really too expensive (and often not smart enough) to run in drone swarms. There are lightshow drones that are really stupid and less of a threat. Most of the big ones you see are like that—they follow pre-prepared flight plans like a dance. As an example, this drone light show had 10,000 drones; they probably cost $200 to $300 each.
And then there are AI-driven drones that can do targeting and are more of a threat. It’s impossible to do a drone swarm without some level of autonomy, as the pilots will crash into each other. Also the cost of the drones here in the U.S. is prohibitive—10,000 drones when they cost $30K each like those made by Anduril would be $300 million for that one swarm. Operation Spiderweb, which damaged or destroyed a third of the Russian Air Force strategic bomber fleet, relied on only 117 drones. A tiny number.
The greatest challenge in a drone swarm would be making one that has insulation (shielding) from high energy CUAS; computer vision for automated targeting, which makes jamming (GPS and control signal) irrelevant; the ability to follow terrain and use clutter for cover; the ability to act in an orchestrated swarm using visual referencing.
As many others have said, in a real shooting contest with a near-peer adversary, the first thing they would do is find all your ultra-expensive, anti-drone high explosive weapons and destroy them with missiles, artillery, multiple drones from different directions, etc.
Israel’s Iron Dome is great, but if Iran launched 10,000 missiles at Israel, then Israel’s Iron Dome would quickly be overwhelmed and run out of missiles.
Can all the components of a drone—wings and hardware, software and chips, consoles, codes, etc. be designed, created, and manufactured in the U.S.? What are the logistical challenges? Can we nearshore some, if not inshore? What about regulations or labor costs in the U.S.?
Yes, they can all be manufactured (to some degree—everything uses components at some level) in the U.S. For sure, NDAA-compliant drones are made using components made around the world apart from in China, just as we are, then assembling in the U.S.
The main challenge is volume and competing with DJI. DJI does such a great job of making drones, all our American capitalist consumer friends are going DJI. This leaves the police and military—and even the police are buying DJI whenever they can right now. If you are selling 300 drones a month, you are in the exquisite platform business, and each of those drones has to be very expensive.
It’s not only the hardware manufacturing challenge, it’s also the software—DJI has hundreds of developers on its firmware teams, and that is leveraged across many drone units. Their object avoidance is better than that of other U.S. drones that cost 10 times as much.
Where does the U.S. stand on using automation for deterrence in schools, on the border, etc?
Nowhere right now. We are working on this problem.
What might be the cost, or how much would a government grant need, to do some drone startups? Where might such grants be spent more efficiently?
I think it would make sense to fund a DJI-killer grant program. States and the federal government have attempted to ban DJI drones, but if there’s no real replacement; it feels like going back to the dark ages for law enforcement. A $28,000 drone cannot replace a $2,000 drone unless there’s a long-term operations and maintenance tail that’s also funded.
I think that such a grant program should have specific criteria—cost of the final system, autonomy, etc. Almost like a pitch contest, where applicants need to convince the government that it will result in a real system for sale commercially, not an expensive science experiment that sits on a shelf.
Who might be the leaders in the field?
Creating successful U.S. champions to counter DJI requires a mix of capability and cost that can only be driven by scale and a desire to ship many units. Drone as First Responder (DFR) programs are valid opportunities for drones, but DFR drones typically cost $100,000 per year per drone, and that’s typically because they are being purchased in small quantities to replace police helicopters. That’s never going to challenge DJI.
Mithril Defense has targeted a cost per drone of $1,695 for an NDAA-compliant, autonomy-capable drone. It is leveraging that in new business models such as school safety that can drive far greater volume and so allow the company to be profitable at far lower per-unit cost than other American drone companies that are focused on niche use cases with very low volume requirements.
This interview has been lightly edited for clarity.