The war in Ukraine is, at once, a past and future conflict. On the one hand, Russia’s invasion resembles World War I, with static frontlines, trenches, and vast areas of no man’s land in which soldiers are quickly killed. On the other hand, the conflict has been defined by innovation and modern technology. There has never before been a war with such broad and effective use of satellites, autonomous systems, AI software, and other commercial gear.

For the United States, these trends should be deeply concerning. After scoring a quick and decisive victory in the 1991 Gulf War, American military planners concluded that future wars would also be short. They assumed that the key to success was no longer mass but large, expensive, high-tech systems. But Ukraine demonstrates the opposite. Wars can still be long conflicts of attrition. Small, poor states can surprise and overwhelm larger, wealthier ones. And success involves learning how to make and use huge numbers of cheap weapons, not simply small numbers of exquisite systems.

Unfortunately, it will be difficult for Washington to change in accordance. The U.S. military is a huge, staid bureaucracy. It has deeply engrained processes and traditions. In fact, some U.S. military leaders have downplayed how applicable Ukraine is to a future conflict in the Pacific. On the website War on the Rocks, the retired general David Barno and the political scientist Nora Bensahel warned that Washington risks “learning the wrong lessons entirely” from Ukraine should the conflict “threaten [the United States’] evolving new doctrine and expensive investments.”

It is true that not every tactical lesson from Ukraine is relevant for future wars. But it is simply undeniable that imagination, adaptability, and affordable, commercially available capabilities will be essential to tomorrow’s conflicts. American defense officials must therefore step out of their comfort zones and start rethinking Washington’s military doctrines. They need to begin envisioning how an adversary’s capabilities could neutralize the United States’ sophisticated defense platforms and tactics. They have to learn how to adapt their systems at the speed of software development, which means within hours and days instead of years. Finally, they must design cheap weapons that American manufacturers can quickly produce and replace, such that total mass—rather than just platform exquisiteness—becomes a significant warfighting factor. Otherwise, when the next conflict arrives, the United States may be shocked and set back, much as Russia was.

ANTICIPATION AND ADAPTATION

The first lesson the United States must internalize is the need for more imagination—specifically, the need to imagine what combinations of new technologies an adversary could employ to neutralize Washington’s advantages. In Ukraine, it is the tools that hadn’t been previously conceived of that have often been the most devastating. Since in a future conflict the United States might be in the position of Russia, Washington must dedicate resources to anticipate how an adversary might use asymmetric tools and tactics to disrupt U.S. operations.

Consider, for instance, the battle for the Black Sea. When the war began, Russia had a large fleet of warships stationed in the waters around Ukraine to blockade the country. Kyiv, meanwhile, had only one, tiny warship. As Russian ships began heading toward Odessa, Ukraine opted to sink the sole warship it had out of fear that Moscow might seize it. But over the next several years, Ukraine compensated for its lack of a traditional navy by using domestically developed missiles, uncrewed surface vessels, aerial drones, and good intelligence to either destroy or disable at least 25 Russian ships—or around a third of Russia’s Black Sea Fleet. That included the Moskva, Russia’s flagship. Ukraine’s campaign ultimately broke Moscow’s blockade, allowing Ukrainian grain to again flow to foreign markets.

The sea is hardly the only place where Kyiv has used creativity to surprise its more sophisticated opponent. In the spring of 2025, after 18 months of planning, Ukraine smuggled 117 first-person-view drones into Russia in trucks positioned near five distant airfields. It then launched them at Russian strategic bombers across several thousand kilometers, damaging at least ten percent—and possibly as much as 30 percent—of the country’s long-range, heavy-payload aircraft. In addition to degrading Russia’s long-range missile strike capabilities, the attack inflicted a psychological blow, demonstrating that no Russian assets were out of reach. The drones Ukraine used cost, at most, a few thousand dollars each. But the damage to Russia’s fleet was $7 billion.

The American military has largely had the same programs and training for the last 30 years.

With these operations and others—including Ukraine’s first underwater drone strike, carried out in December against a Russian submarine—Kyiv consistently thought outside the box because it did not have the resources of a superpower. It had to be creative and improvise, and this led it to use commercial technologies in ways that offset some of Russia’s traditional strengths. Technologies that are advancing today will give states even more opportunities to gain asymmetric advantages in the future.

Although anticipating new forms of attack is critical, not everything can be predicted before a conflict begins. Because adversaries conceive of new ideas with every technological advance, the United States must also adapt faster. The importance of rapid adaptation is the second lesson Washington must learn. Although on display in many dimensions of the Ukraine conflict, rapid adaptation is most evident in the drone-counterdrone innovation cycle. At the beginning of the war, Ukraine pioneered using disposable, uncrewed vehicles to attack enemy positions. But by the end of 2022, Russia had learned to deploy jammers along the frontlines to disrupt Ukrainian signals when the drones got within several kilometers of their targets. Kyiv then adapted again by buying and using small, commercial, off-the-shelf computer vision models, which allowed frontline operators to lock drones onto targets outside the jamming range. But then, the Russians expanded the ranges of their jammers and deployed smokescreens to protect Ukraine’s targets. This cycle has since continued and even sped up. Today, tit-for-tat tactical improvements happen so fast that previous innovations can become obsolete within three weeks. Although war tactics have always adjusted in real time, this pace of frontline, software adaption is unprecedented.

Other countries are also innovating by taking advantage of rapid advances in commercially available technology. In November, for instance, the AI lab Anthropic released a report detailing how, for the first time, a Chinese state-affiliated hacking group orchestrated a powerful cyber-espionage operation against large U.S. enterprises almost entirely using Anthropic’s Claude AI tools. Those tools have become publicly available only within the last year, and Anthropic, Google, OpenAI, and others have released even more powerful tools in the months since. The fact that AI capabilities improve so quickly means that state and nonstate actors alike might gain access to new ways to hack, build bioweapons, and orchestrate swarms of autonomous systems.

The final lesson Washington must accept is the need to design low-cost systems, which allows for building at scale. Here, too, drones serve as an encapsulation. Ukraine’s use of large numbers of low-cost drones has become a hallmark of its warfighting approach. Even with Russian jamming, drones built for a few hundred dollars apiece allow Ukraine to offset Russian advantages in artillery, armor, and massed fires. They can deliver precision effects that previously required far more expensive munitions, such as precision-guided Excalibur shells, which can cost $100,000 apiece. Drones, in other words, have provided a more flexible form of precision firepower that can be flown by small teams close to the frontline.

For the United States, a low-cost design approach would be most effective as part of a strategy that combined a small number of high-end, exquisite platforms with large numbers of lower-cost, simpler, more “attritable,” or easily expendable, ones. Washington has already invested in exquisite platforms, but it is only in the early stages of procuring attritable assets, and it needs to catch up.

The United States can learn much from Ukraine about how to build and buy such systems. But it should not simply follow Kyiv’s template. Ukrainian drone production is still reliant on Chinese components, such as motors, flight controllers, navigation cameras, batteries, and thermal sensors. The irony of having Western and Ukrainian aid channeled through Chinese supply chains is that it ultimately supports Russian and Iranian drone production, given that Beijing is also an indispensable arms dealer to both those countries. This fact highlights why access to cheap, non-Chinese drone components is essential for the United States and its allies.

The U.S. military is still the most sophisticated armed force in the world. But it will struggle to adjust to these lessons—which go beyond specific technologies or tactics. There are several reasons why. For starters, the American military is an enormous bureaucracy, and it has largely had the same programs and training for the last 30 years. Such entities are never easy to change.

The country’s defense preparations provide a case in point. American defense officials have spent decades preparing for obvious attacks, such as strikes that could blind U.S. satellites or cyberattacks that could disrupt the country’s digital infrastructure. These plans are necessary, but American officials have not prepared for less apparent lines of attack or even considered what such lines of attack might be. This makes the American military vulnerable in the same way that Russia’s was when Ukraine—without a regular navy—attacked ships that Moscow assumed were out of reach.

The United States military will also struggle to adapt in part because of its commitment to exquisite systems, which are nearly impossible to quickly upgrade or build at a speed fast enough to replace losses. The F-35 fighter jet, for example, was in development for roughly 20 years. Aircraft carriers take ten years to construct. The B-21 bomber has been in development since the early 2010s, but it will not reach full operational capability for several more years. The first contracts for the navy’s Extra Large Uncrewed Undersea Vehicle were awarded in 2010; Boeing, the primary contractor, has still not delivered an operational vessel. Even innovations that the Pentagon celebrates for their speed proceed at a glacial pace by Ukrainian standards. When Iraqi insurgents began blowing up U.S. Humvees with improvised explosive devices in late 2003, the American military rapidly rushed to develop mine-resistant vehicles. The MRAP, as it was called, still took 20 months to produce.

These exquisite capabilities don’t just have long delivery cycles. They also have enormous budgets, with cost overruns that have become commonplace as delivery timelines stretch on. These overruns are further exacerbated by cost-plus contracting, in which suppliers are paid a fixed percentage of the total costs as profit margin. As a result, a single F-35 costs the government $80 million, and a Ford-class aircraft carrier costs $13 billion. These exquisite platforms are so expensive that the military cannot afford to lose them. The Pentagon could buy 13 million drones for the cost of an aircraft carrier, of which the U.S. Navy has 11 (and it plans to build more), which would supply nearly 100 drones for every U.S. Army and Marine infantryman.

Even if Washington began spending more on low-cost systems, it would be starting from a difficult position. Over the last 15 years, China has supplanted the United States as the global manufacturing powerhouse. It is able to scale up the production of ships, missiles, drones, and electronics far faster and at a lower cost than the United States can, which could give Beijing a decisive advantage in a prolonged conflict between the two countries. China’s shipbuilding capacity, for example, is 200 times that of the United States.

THE QUALITY OF QUANTITY

Washington does have experience in both envisioning how its adversaries can use asymmetric tactics and ramping up production for war. In the lead-up to World War II, the United States war-gamed so effectively that Admiral Chester Nimitz later commented that the only surprise was the kamikaze attacks that happened late in the conflict (although Washington also failed to anticipate Pearl Harbor). The United States also dramatically accelerated military manufacturing during the conflict, producing more than 5,000 ships and 300,000 aircraft. Quantity, officials discovered, does have a quality all its own.

The United States should learn from this history. Defense planners can start by mimicking their interwar predecessors and carrying out more war games. They should move beyond linear tabletop exercises with scripted vignettes of sequential moves to a more realistic environment that simulates the fog of war. These games should force planners to make rapid decisions with incomplete information, degraded command-and-control systems, constrained logistics, and political limits—just as Russia and Ukraine have to.

The Pentagon should also hire outside firms to engage in red teaming—or to think up situations it might otherwise miss. It should reconstitute the Office of Net Assessment, which provided independent, long-range military assessments that shaped U.S. strategy and investments but was disbanded without explanation in 2025. The Pentagon should require the Joint Force planning staff—planners from each military service who serve on the chairman of the Joint Chiefs’ staff—to brainstorm asymmetric strategies that U.S. adversaries might deploy. Military planners should develop ways for U.S. forces to keep fighting if opposing militaries degrade American communications, intelligence, and other systems—just as U.S. forces are planning to degrade adversaries’ systems. AI allows planners and strategists to run many more simulations of these scenarios than they could in the past.

The U.S. armed forces needs hardware that helps them adapt faster.

To prepare for the future, however, the Pentagon shouldn’t only have strategists do red-team drills. Soldiers themselves must do them—and treat each one as a laboratory for trying new tactics. American forces need frequent, realistic field exercises that stress test units in contested environments to improve their improvisational skills. Troops should also undergo exercises that are longer than the standard two weeks, so they can learn how to better resupply and overcome logistical hurdles. During these exercises, U.S. officials should mandate that the troops who are role-playing as the adversary develop and employ asymmetric tactics.

The army’s “transformation in contact” exercise, carried out this fall through the Joint Pacific Multinational Readiness Center, was a good step in this direction. Although this exercise was not long enough to fully test the U.S. military’s resupply capabilities, it did begin to emphasize the logistical and technological challenges that U.S. troops could face in the wars of the future. It also included all the services, not just the army, plus military personnel from seven allied countries. Many of its 75 experiments involved new technologies. Critically, the exercise placed soldiers in situations in which events did not proceed according to plan.

Exercises are essential to improving the military’s adaptability and forcing it to think through new scenarios. But to prepare for Ukraine-like wars, the U.S. armed forces also need new hardware that helps them adapt faster. The U.S. military, for example, must design and procure modular systems—or systems made up of commercially available components that can be easily replaced and upgraded. In theory, the Pentagon already requires a “Modular Open Systems Approach” for any new weapons program, when feasible. But in practice, the military has carved out overly broad exceptions because producing modular systems can initially be more costly and time-consuming than producing nonmodular ones. In the long run, however, modular designs save both money and time by reducing the costs of sustaining older systems. The military, for example, maintains older ships, tanks, and planes for decades. As a result, 70 percent or more of the lifetime cost of these systems is incurred after initial procurement.

Russia’s invasion of Ukraine was meant to last three days

Expanding productive capacity and creating supply chains ready to produce enough war materiel—especially drones and munitions—will almost certainly be a lengthy endeavor. The United States, after all, is starting far behind. Today, the U.S. Navy and U.S. Air Force have only 300 ships and 5,000 aircraft, respectively, a shadow of what they had during World War II. The United States is set to produce less than a tenth as many drones as Russia and Ukraine will each make this year. Washington will thus have to make substantial investments in both domestic and allied-based drone component companies. (It must also stockpile drones, drone parts, and munitions to avoid running out of these goods when wars inevitably last longer than expected.) But with a $900 billion defense budget, Washington has the money to catch up. Today, only 20 percent of this budget goes to buying ships, tanks, planes, software, or any other kind of kit for troops. The government could up that percentage by reallocating some of the 17 percent of the budget currently dedicated to developing new exquisite platforms, which, as Ukraine has shown, are likely vulnerable to asymmetric attacks by the time they are fielded.

The United States can also speed up by adopting advanced manufacturing technologies, such as 3D printing, which better allow for replenishment and upgrades in the field. The country is already home to several startups that are building 3D-printed drone assembly systems that fit entirely within a 20-foot shipping container and reactors that can make hydrogen fuel on site with just a metal catalyst and seawater. These systems can help ensure that frontline soldiers and commanders get what they need quickly rather than waiting for weeks for supplies to come in. A damaged F-35, for example, might take months to fix, if it can be fixed at all, as the military waits for manufacturers to produce a new custom component. A modular drone can be repaired in a matter of days with parts sitting on a shelf, or less time if the parts can be printed.

Russia’s invasion of Ukraine was meant to last three days; instead, it has dragged on for four years. The United States needs to study why Moscow, previously thought of as a leading military power, failed and why Kyiv has performed so well. Washington must learn how weaker opponents can surprise stronger ones, prepare to iterate rapidly, and start producing large quantities of low-cost capabilities. These new paradigms of conflict are not easily or quickly internalized, which is exactly why adopting them requires urgency and large-scale focus.

You are reading a free article

Subscribe to Foreign Affairs to get unlimited access.

Paywall-free reading of new articles and over a century of archives

Six issues a year in print and online, plus audio articles

Unlock access to the Foreign Affairs app for reading on the go

Already a subscriber? Sign In

© Foreign Affairs