An Industrial Policy for Drones

Turn any article into a podcast. Upgrade now to start listening. Members can share articles with friends & family to bypass the paywall. You’re reading Dispatch Markets, a weekly dive into the forces driving economic growth—and those holding it back—featuring Scott Lincicome, Kyla Scanlon, Karl Smith, Marian Tupy, and Adam Ozimek. As someone who straddles the line between critic and proponent of industrial policy, my position is that it can work but it's very hard to get right. It requires good policy design, but you also need the fundamental economics on your side. Industrial policy, for those new to the topic, is a state action meant to expand a particular industry. In general, such policy will be considered a success if an early and cost-effective push leads to sustainable and self-supporting commercial growth. This means no bridges to nowhere, and no bottomless funding pits. Let’s start with one success and one failure. A Department of Energy loan to Tesla was a successful industrial policy for pushing the electric vehicle industry forward because it helped a critical company survive. But it could succeed only because electric vehicle technology was on the verge of commercial viability. In contrast, a 1980s government program to develop a synthetic fuels industry failed because synthetic fuels, at the end of the day, did not make economic sense in the long run even with a billion-dollar boost. It can be extremely hard to know in advance whether the economics are on the side of a particular industrial policy, and thus a large dose of humility is in order. But I believe there is a case that drones at least have a good shot at being the target of successful industrial policy. Betting on Drones Illustration by Noah Hickey/The Dispatch (Photos via Getty Images). Let me begin with a simple sketch of my policy idea that I would like to pitch here for the very first time: Drone to the Top. The federal government will commit to spending a fixed amount annually, for 10 years, on drone swarms for organizations that want to use them for the public good. Organizations will submit short proposals for how they’d use the swarms in innovative ways that generate public value. Every year the federal government pre-approves a list of drone makers whose supply chains are in the U.S. or selected allies. The federal government picks winning organizations who then get to choose which drone maker suits their particular needs, and the government buys a drone swarm for them. The federal government advises winning organizations on implementation and then publishes annual reports of what was learned from the year’s case studies. The result of this is organizations across the U.S. with access to drone swarms for purposes of the public good, and drone industries located in the U.S. and allied nations getting what is called an “advance market commitment” that lets them know there will be demand. Why drones, though? You may be familiar with drone swarms from your local summer fireworks display. The technology continues to progress, with a recent world record of 22,580 simultaneous drones controlled by a single computer. Such swarms can be an entertaining spectacle, but also have potential real-world uses. In contrast with single drones, researchers argue that “drone swarms can perform complex tasks by utilizing decentralized operation and data processing.” These capabilities are already being used on the battlefield, but these researchers argue drone swarms would also be useful in applications like “monitoring rapidly evolving ocean phenomena … and target detection and tracking.” This suggests a wide range of potential non-defenses, which I will say more on below. But let me begin by making the case that the economics line up for supporting industrial policy in this case. Drones matter. Start with the national security case. Even the most strident opponents of industrial policy will usually agree that a national security argument can potentially overcome objections. It is true that national security is often a fig leaf, rationalizing protectionism for almost any good. Even peanut butter. But the war in Ukraine has made clear that drones are playing a major role in 21st-century warfare. Since the war began, both sides have utilized drones offensively and defensively. It is no coincidence that Ukraine is the first country to have an entire branch of its armed forces dedicated to unmanned systems, which includes drones. In 2025, Ukraine’s drone pilots were flying 10,000 combat missions a day, and the country’s chief military commander estimated that they accounted for 60 percent of Russian casualties. On the other side, Russia is deploying thousands of drones a day as well. Drones have also played a large role in the war in the Middle East, with Iran using drones to target both military and civilian infrastructure. The national security concerns here are real. Drone policy in the real world has never mattered more, as drone production has become an important lever in modern war. In just a few years, Ukraine has gone from a handful of drone manufacturers to being called “a drone superpower” by the Atlantic Council, capable of producing millions of drones a year. Russia is working to increase its kamikaze drone capacity to reduce reliance on China, reportedly aiming to build 130,000 such drones annually by 2030, and 350,000 by 2035. In terms of first-person view (FPV) drones, which are smaller and often resemble the small quadcopter-type drones that hobbyists use, Ukraine’s top military commander alleges Russia is aiming for an annual production of 7.3 million. The U.S. is also focused on improving our domestic drone capacity. For example, the Defense Department is sponsoring a Drone Dominance contest for commercial manufacturers to compete in a series of tests, with winners ultimately providing 300,000 drones by 2027. The goal is to provide “a steady demand signal and a commercial competition that leverages private capital.” Importantly for industrial policy, drones are not like tanks. A lot of innovation here is happening in the commercial sector, not at larger traditional defense manufacturers. This means that policy also needs to target the commercial sector. As Kateryna Bondar at the Center for Strategic and International Studies argues: Today, most cutting-edge technologies—whether in software, AI, or unmanned systems—are originally developed in the commercial sector, targeted at consumer or B2B markets, and shaped by open competition and rapid iteration, not by traditional defense firms. The true drivers of innovation in unmanned and software-centric technologies lie within the commercial sector, including startups and venture capital firms, which operate outside the rigid, decades-old structures of the traditional defense ecosystem. For the U.S. this policy is not just about building one or two particular drones onshore. And some drones, like large military drones, will not be available or applicable to local organizations in need of drone swarms. Instead the goal is boosting the entire commercial drone industry and its suppliers to make the “electric tech stack,” with its wide defense uses, more friend-shored and onshored. Two-sided industrial policy. Normally, an industrial policy is rationalized by pointing to some way that the policy makes the industry more productive. There’s a variety of ways this can happen, including economies of scale and learning by doing. What makes drones, to me, a particularly compelling case for industrial policy is the potential for benefits on both sides of the market. Let’s discuss them in turn. Producer side: Drone manufacturers who are successful at Drone to the Top will receive advance commitments to future demand, helping them invest to produce at scale. In other words, we may get economies of scale that would be absent without the program. This demand can help producers reach an economically sustainable scale and help overcome coordination problems in the supply chain. Drone companies need suppliers to exist, but suppliers need drone companies to exist. Communicating to everyone that demand will be there helps them to coordinate, whereas otherwise everyone may sit back and wait for others to invest first. By allowing the applying organizations, instead of a centralized bureaucratic decision maker, to choose the drones, Drone to the Top also ensures drone makers will be in close contact with direct users of their goods and have the opportunity to learn from them. This kind of “learning by doing” is one of the lessons of Ukrainian drone policy. Ukraine has authorized 700 military units to procure drones (among other technologies) directly from approved commercial providers, which allows troops on the ground to clearly communicate their needs to drone producers In a 2025 report based on interviews with more than 50 drone experts and stakeholders inside Ukraine and beyond, Bondar at CSIS argues this procurement approach reduces the role of centralized bureaucracy and makes acquisition more “tailored to the real-time needs of frontline operations.” In contrast to procurement policies where “rigid technical specifications” are issued, the Ukrainian military “now articulates their needs as operational problems communicated directly by end users.” To that end, CSIS recommends the U.S. military take the following approach for drones: Instead of relying on rigid engineering requirements set far in advance, the U.S. military should define its needs through operational problem statements developed directly by those who will deploy the technology in the field. Drone to the Top would work similarly. User side: It is not just the producers who stand to benefit from learning by doing. Commercial drones cheap enough for swarms represent a newer technology, and we are still in the early stages of learning what they can do. Drone to the Top would provide organizations with subsidized access to drone swarms they wouldn’t otherwise acquire. This allows potential users on the ground to experiment and learn, but the policy would also codify the process of learning from users and sharing this knowledge. For example, police departments across the country could learn by example if a police department wins Drones to the Top and issues a report on implementation lessons. What will these organizations do with their drone swarms? Part of the value of this contest is elevating ideas that we cannot come up with today. Drones are a newish technology, and drone swarms are even newer. But here are some potential examples of both the user and the use case: A local search and rescue wants to use drone swarms armed with thermal cameras to find lost or injured hikers. A fire department wants to develop rapid-response swarms to put eyes on a fire, which would help estimate the scale of the fire, allow firefighters to plan an appropriate response, and maybe even do fire suppression. A police department wants to use drones to track criminals and those evading police. University geologists want to map and scan for mineral deposits. A university agricultural department wants to experiment with distributing fertilizer. A Coast Guard unit wants to use drones for ocean search and rescue. As these organizations learn how to use drone swarms for the public benefit, what they learn will be iterated upon, codified, and shared by the federal government. The case for knowledge spillovers, another industrial policy justification, is clear here. A lesson from history. Between national security and benefits on both sides of the market, the rationale for industrial policy is coming from three directions here. And usefully, all three are mutually supportive, with a shared goal of spurring commercial production. We have seen a similar case before with the semiconductor industry. Walking through the history shows why industrial policy might work here, and why it doesn’t work other times. The first consumer use of an integrated circuit (IC) was in a Zenith hearing aid in 1964. Up to that point, the U.S. government was responsible for almost all purchases of such circuits. This included the onboard computers for the Apollo program and the guidance system for the Minuteman II intercontinental ballistic missile. These programs undoubtedly boosted the industry at a stage when semiconductors were seen as expensive and risky. As economist David Mowery argued, “The prospect of large procurement contracts appears to have operated similarly to a prize, leading Texas Instruments to invest its own funds in the development of a product that met military requirements.” But while many ask what the government can do for industry, it is important to recognize what industry did for the government. It was not simply that the federal government saw an opportunity to spur growth in a nascent technology. Instead, it had serious needs that were “acute, even existential.” The problem was the “tyranny of numbers.” Before integrated circuits, computers were made up of individual components like transistors, resistors, diodes, and capacitors all wired together. As the 1950s progressed, electronics became increasingly complex. The components were all independent, so building a computer required extraordinary amounts of wiring and soldering. The Control Data CDC 1604, the most powerful computer at the time of its release in 1959, contained 25,000 transistors, 100,000 diodes, and hundreds of thousands of other components. New Navy aircraft carriers had 350,000 electronic components, with millions of hand-soldered connections. With so many individual connections, the odds of failure were high, and these systems were fragile. These systems were also large and heavy, which is a problem for missiles and spaceships where every pound counts. Both government and industry searched for a solution to the “tyranny of numbers.” Indeed, all three military services pursued separate independent technological solutions (Navy “thin film,” Army “micro-module,” and Air Force “molecular electronics”), none of which worked. The solution came to researchers at Texas Instruments and Fairchild Semiconductor, who figured out how to put multiple electronic components onto a single chip. Integrated circuits solved the tyranny of numbers and addressed real needs for the government. But integrated circuits were not just smaller and lighter, they also were durable and reliable. There were fewer failure points. As IC co-inventor and later co-founder of Intel Bob Noyce wrote in 1964, “Reliability is the most important single factor.” This was essential for the Minuteman and Apollo use cases: Rocketing through the air at high speed, either as a missile or space ship, necessitates very high levels of durability and reliability. The case of integrated circuits shows the industry investing in response to the existence of government demand for a real, concrete need, and this need happened to align with the direction of future market demand. In other words, the government did not buy ICs for the purpose of spurring an infant industry. Officials bought them because they solved real problems. There were benefits on both sides. Drone to the Top is designed to do the same thing for drone swarms. I don’t think we should embrace industrial policy willy-nilly, nor do I think the median industrial policy is going to be successful. But it is very helpful when the economics suggest there are benefits on both sides, and when you can take a more bottom-up versus top-down approach. This is a sketch of a policy design; it is not ready to be shipped yet. But I think it is worth considering and developing further. Markets FTW Owen Zidar and Eric Zwick have a new book out called The Everywhere Millionaire about who is a millionaire in the U.S. economy. In a Wall Street Journal piece summarizing their work, they argue there are 3 million everywhere millionaires, worth $65 trillion combined. I suppose one could find such inequality discouraging, but as Zwick and Zidar find, the path is open to those with the hustle to become entrepreneurs: “Indeed, the typical path to $10 million and up comes from owning a company. That path is open in every town, in unglamorous industries, to people without top test scores, fancy degrees or rich parents.” I see this as preferable to Harvard and high finance being the only route to riches. I like this content creator named Chris Koerner, who is all about finding and documenting these sorts of everyday businesses that people can pursue as a path to high incomes and independence. If you want to be an everywhere millionaire, the path is out there. Chart of the Week The rural population bounce, large urban population dip. Worth Your Time Someone ported Doom to the Vectrex videogame system. Fiat has a $15,000 car, but is anyone going to buy this? A new dataset on Commuter Market Access from Mercatus. An argument to let the Chinese car imports in. It would be a very bad idea to put a huge tax on foreign students working but they’re trying to do it. Disclaimer: The opinions expressed above do not necessarily reflect those of the presenting sponsor. Adam Ozimek is chief economist at the Economic Innovation Group, where his work focuses on labor markets, immigration, regional growth, and the policies that shape American economic dynamism. He holds a Ph.D. in economics and previously served as chief economist at Upwork and as a senior economist at Moody’s Analytics.