GREG LITTLE is a senior counselor at Palantir Technologies.
IAN CRONE is a member of the technical staff at Palantir Technologies.
In 1980, Detroit was in a fight for its life. Two oil shocks had pushed American buyers toward fuel-efficient Japanese imports, Chrysler had just been rescued by federal loan guarantees, and the Big Three were watching their market share erode year after year. Executives no longer dismissed Toyota as a maker of inexpensive economy cars in a different category; they acknowledged Japan as a serious competitor to their core business. What they did not acknowledge—what few of them even perceived—was the Toyota Production System. Detroit believed it was losing on fuel economy, labor costs, and exchange rates: problems that better products and better terms could presumably fix. To some extent that was true, but Toyota understood that the real competition was system against system.
For years, American automakers focused on horsepower, styling, dealer networks, and consumer preference. Toyota was building something more consequential: a production system that connected workers, suppliers, engineers, logistics, quality control, and leadership into a continuously learning enterprise.
The Toyota Production System was more than a collection of efficiency techniques. Its two pillars—just-in-time production and jidoka, or “automation with a human touch”—established a fundamentally different way of seeing and managing the enterprise.
Jidoka is the more remarkable of the two principles because it hard-wired human-like judgment into the machines themselves. The idea predates the automobile business entirely: Sakichi Toyoda, the company’s founder, invented an automatic loom that stopped instantly if a warp or weft thread snapped. Because the machine could distinguish normal from abnormal conditions, defects were identified and the process stopped before they turned into expensive mistakes. In Toyota’s plants, any worker could stop the production line upon discovering a quality problem—and so could the machines.
The Toyota Production System extended beyond the factory floor. Suppliers became integrated partners rather than transactional vendors, while data was used to prevent tomorrow’s defects rather than merely explain yesterday’s failures. Every production run generated feedback, every problem became an opportunity to improve, and every improvement strengthened the broader system.
By the time Detroit understood what was happening, Toyota’s advantage was bigger than any individual vehicle or factory. It was embedded in the relationships, processes, information flows, management practices, and learning loops that connected the enterprise. American manufacturers eventually adopted many of Toyota’s methods, but they spent decades trying to catch up to a system that had evolved under the surface while they remained focused on the product.
Something similar is happening today with America’s defense industrial base, a system that is fragmented, expensive, and slow to learn during a period of rapid change. America cannot solve the problems with this system merely by routing more money through the same fragmented processes. Just as Detroit could not defeat Toyota by adding another shift to an inferior production system, the Department of War cannot create a twenty-first-century arsenal by applying more funding, oversight, and reporting requirements to a twentieth-century operating model. What it needs is a cutting-edge industrial operating system that gets better and smarter over time.
Going to War With the System You Have
Toyota’s lesson is fundamentally about how to win an industrial competition, and the lesson applies to more than cars. Wars are also decided by industrial systems.
America’s logistical and production juggernaut—what we now nostalgically call “the Arsenal of Democracy”—won World War II. In the present day, Ukraine is holding its own against Russia with a highly adaptive industrial system capable of developing and scaling production of drones to counter the latest threats encountered on the front line. Now Epic Fury has exposed shortcomings in America’s industrial system, which must be able not only to enter a fight with overwhelming force but to sustain the fight over the long term.
The lessons of these conflicts are vital to the United States’s competition with China, a challenge with far higher stakes than the Japanese auto industry’s challenge to Detroit in the 20th century. Much of the debate focuses on visible measures of military power: ships launched, aircraft delivered, missiles fielded, budgets approved. Those measures are important, but they are the outputs of an industrial system. More important than simply counting missiles is whether the system producing them can learn, adapt, scale, repair, and replenish faster than its competitor.
China brings a formidable industrial system to the fight. This system is neither perfectly coordinated nor perfectly efficient, but it is well suited for its strategic goals. It only needs to convert its advantages in commercial manufacturing, shipbuilding, electronics, critical materials, infrastructure, and state-supported investment into military capability faster than the United States can overcome its own fragmentation. Its military-civil fusion approach has intentionally reduced barriers between commercial and national-security industries, and its dual-use shipbuilding ecosystem demonstrates the strategic value created when commercial orders sustain the infrastructure, suppliers, workers, and engineering capabilities that also support naval construction and repair. Analyses of China’s naval buildup and dominance of global shipbuilding demonstrate why industrial scale, supplier depth, workforce, and production experience matter as much as the design of any individual vessel.
It should be no surprise that Chinese EVs are taking the world by storm just like Japanese autos did a half-century ago: China is operating an industrial system that is large, well coordinated, adaptive, and inherently dual use. The West is contending with the China Production System—and we need to overhaul our own system, incorporating our own strengths and values, to prevail.
The United States still possesses extraordinary advantages: exceptionally capable submarines, aircraft, missile defenses, satellites, sensors, and software; world-class engineers; deep capital markets; leading research institutions; innovative commercial companies; and an unparalleled network of allies. The problem is that these strengths are treated as a collection of programs and discrete things rather than as a system that adapts and improves over time.
The Pentagon’s first National Defense Industrial Strategy recognized the importance of resilient supply chains, workforce readiness, flexible acquisition, and economic deterrence, and its subsequent implementation plan assigned responsibilities, resources, and metrics to those priorities. The 2026 National Defense Strategy similarly calls for rebuilding America’s defense industry and restoring its role as the world’s premier arsenal. The strategic direction is becoming clear; turning strategy into coordinated execution remains the harder task.
Program offices optimize individual acquisitions. Prime contractors optimize contractual performance. Suppliers optimize production schedules. Logistics organizations optimize sustainment. Each institution may perform its assigned mission well while the ecosystem as a whole underperforms. A recent Government Accountability Office assessment found the Department’s supply-chain visibility initiatives uncoordinated, limited in scope, and unable to provide sufficient insight into lower-tier suppliers, raw materials, and components.
Consider a missile program that receives billions of dollars in additional funding. Beyond a certain point, money alone will not create more interceptors. The program may depend on a rocket-motor supplier that also supports several other weapons, a sub-tier manufacturer producing specialized insulation or nozzles, a limited testing facility, or a workforce that takes years to develop. The government’s recent investments to expand solid-rocket-motor component production illustrate how a single sub-tier component can constrain an entire portfolio of weapons, and previous assessments of competition within the defense industrial base have documented significant supplier consolidation in exactly these areas. Additional funding may simply send more demand toward the same bottleneck without increasing output—and few leaders can see the shared constraint across the portfolio.
Assembling the War Machine
That example reveals the central problem: leaders lack the information needed to determine which constraint to address first, which supplier to expand, which alternative source to qualify, or which program should receive scarce capacity. They cannot consistently see how a disruption in energetics, semiconductors, castings, critical minerals, test infrastructure, or skilled labor will ripple across multiple weapons over time.
For decades, the Department of War has invested in command-and-control systems that help leaders understand the battlefield: commanders can see the location and status of ships, aircraft, formations, sensors, and threats, and integrate that information to accelerate decisions across domains. Yet the industrial ecosystem that produces, repairs, sustains, and replenishes those forces is still managed through fragmented data, disconnected workflows, periodic reports, and lengthy coordination processes.
The comparison with automobile manufacturing has limits. Defense systems involve classified requirements, low production volumes, exacting standards, specialized materials, congressional funding rules, and threats that evolve during development. The lesson is not necessarily that missiles should be built like Corollas; it is that both enterprises depend on supplier integration, quality, visibility, flow, feedback, and continuous learning—and the more complex the product, the more consequential the operating system behind it.
Artificial intelligence and modern software now offer the United States an opportunity to create that operating system across the defense ecosystem. Research on AI as a general-purpose military technology argues that its most profound strategic effects may emerge from broad productivity gains and spillovers across an industrial base rather than any particular tool or weapon. This is jidoka at national scale: the human-like judgment Sakichi Toyoda built into a single loom can now be imbued into thousands of machines, workflows, and decisions across the enterprise—sensing abnormality, stopping defects before they propagate, and forcing problems into the open where they can be fixed. Program leaders, suppliers, manufacturers, depots, combatant commands, and policymakers could operate from a continuously updated picture connecting operational demand with inventory, production capacity, supplier risk, workforce, and delivery timelines. AI could identify constraints before schedules slip, expose shared dependencies across programs, and establish faster learning loops between operators, engineers, factories, and suppliers. The objective is not to automate every decision, but to give human leaders a more complete understanding of the choices, dependencies, and consequences before them.
This is the idea behind what we call War Machine. The concept should not be confused with another dashboard, database, or acquisition program. It is the recognition that industrial power has become a command-and-control problem. What Toyota accomplished across its production enterprise, the United States must now accomplish across a far larger and more complicated network of government organizations, defense primes, commercial companies, suppliers, depots, shipyards, laboratories, and operational commands.
War Machine is valuable not merely because it shows leaders the industrial base, but because it enables better decisions across programs, suppliers, money, time, and risk: which bottleneck to attack first, which supplier requires intervention, where additional capacity will matter most, when consumption may exceed replenishment, and which investment will produce the greatest combat effect per dollar and per month. The objective is coordinated action from the factory to the foxhole.
The United States must connect operational demand to industrial capacity, manage constraints across portfolios rather than within individual programs, and establish a shared operating picture across government and industry. It must build learning loops that carry information from operations to engineers, factories, and suppliers, and back to the force—and it must measure success through production, delivery, readiness, and operational outcomes, not money obligated, contracts awarded, or activities completed.
Detroit’s mistake in the last century was believing it was competing car against car when it was actually competing system against system. America must not make the same mistake in national security. The competition with China is more than a race to design the best weapons; it is a race to build the best system for producing, adapting, repairing, replenishing, and sustaining them.
