Joseph Jewell, the assistant defense secretary for science and technology, says the U.S. industrial base must learn to field small, cheap, AI enabled weapons the way Kyiv has against Russia.
The Pentagon's top research official is making a structural argument about how the United States builds weapons, and the proof point he keeps returning to is Ukraine's drone industry. Joseph Jewell, the assistant defense secretary for science and technology, told the Defense One Tech Summit in Arlington, Virginia, on Tuesday that the United States must learn to field capabilities the way Kyiv has: invented, produced, and deployed in volume, on a timetable set by the war itself.
Jewell's frame is not a coffee anecdote. It is an indictment of the U.S. procurement model. The U.S. industrial base, he argued, must adopt the wartime logic that let Ukraine's homegrown drone sector "spring up almost overnight" under combat pressure, and then carry that logic into peacetime competition with China. The U.S. can do it, he said, and at a scale Ukraine could not.
The operational evidence he cites is the Black Sea. Ukraine has, in Jewell's telling, pushed the Russian Navy out of the fight using "a lot of" small, "relatively undetectable" weapons, an effect that, in his phrase, "essentially did not exist when the war began." The next step, he said, is autonomy. "I think the natural evolution is going to be a hundred or a thousand drones controlled by AI," Jewell said. Many Ukrainian FPVs, or first-person-view kamikaze drones, are still manually controlled. The transition to swarms is what Jewell is asking the U.S. industrial base to prepare for.
The levers are concrete. The first is a program Defense Undersecretary for Research and Engineering Emil Michael announced in January: a no-fee licensing arrangement for roughly 500 Pentagon-held patents. The first no-fee license was issued last month. As of mid-June, the Pentagon has signed out 14 patents, issued one paid exclusivity license, has 36 more pending, and is reviewing 145 additional applications, according to Defense One. The department holds tens of thousands of patents and collects roughly $20 million a year from them. The reform is small in dollar terms and large in cultural ones. It is meant to lower the barrier for small companies and nontraditional defense vendors to build on Pentagon-funded research, the same vendor pool that produced Ukraine's drone sector.
The second lever is manufacturing. BioMADE, a Defense Department Manufacturing Innovation Institute, has developed a bioengineered thermal coating to help drones obscure their heat signatures. In the Pacific, Marines have field-3D-printed shaped charges, focused explosive charges that direct blast energy into a narrow jet used to penetrate armor, using endemic materials: plastic water bottles, crushed volcanic rock, coconut husks, and coffee grounds. All detonated. The printed charge, per Jewell, showed 25% better focusing characteristics than conventionally manufactured high explosives. Volcanic rock fragments performed best of the field-mixed options.
Jewell's forward picture is containerized. He described production facilities the size of a CONEX box, the standard military shipping container, that could be moved into the Indo-Pacific and produce shaped charges, biodiesel, and jet fuel in place. The point is not the technology itself. The point is moving the production line to the theater, the same way Ukraine moved drone production to its front lines.
The honest reading is that none of this is policy yet. Jewell is the Pentagon's research chief, not the acquisition chief. The exquisite-platform programs that buy a smaller number of very capable, very expensive systems still define most of the U.S. weapons budget, and the cultural and procurement friction between those programs and a mass-produced, AI-enabled drone doctrine is the friction Jewell is trying to name. His case is that the U.S. cannot wait for the next war to invent what it should already be building.
The watch item is whether the patent holiday, the bioengineered coatings, and the CONEX-box fabrication line move from summit rhetoric to program-of-record status, the formal designation that locks in funding and acquisition priority. That is the gap between a stated direction and a procurement system that can actually deliver on it.