The Pentagon Must Be a Customer, Not a Captive
How the defense-tech revolution could strengthen America at the edge while leaving its most important industrial monopolies untouched—and what Congress should do about it
The arrival of a new generation of defense companies has created a welcome sense that the United States may finally be escaping the procurement habits that have slowed weapons development for decades. Firms such as Anduril, General Atomics, Saronic and others promise shorter development cycles, greater reliance on private capital, software-driven iteration and products designed for manufacture rather than for an endless sequence of government-funded demonstrations. Their emergence is real, useful and overdue.
It is not, however, sufficient.
The most dangerous outcome would not be the failure of these new firms. It would be their success in a limited part of the market while the old industrial structure becomes still more entrenched where failure matters most. New entrants could win drones, autonomous vessels, sensors and software while the traditional primes consolidate their control over submarines, aircraft carriers, nuclear propulsion, large combatants and the specialized supply chains beneath them. America would then possess innovation at the edge and monopoly at the core: an agile layer of relatively inexpensive systems resting on an increasingly brittle foundation of late, costly and irreplaceable platforms.
That possibility changes the reform question. The issue is no longer simply how the Department of Defense can purchase innovative products more quickly. It is how the federal government can prevent the defense-tech revolution from becoming a pressure-release valve that leaves the most consequential monopolies intact.
The governing principle should be straightforward:
DoD should be the customer. It should not also be the captive customer, industry regulator, bailout authority and final judge of contractor performance.
What the Second World War Actually Proved
The American mobilization of the Second World War is sometimes described too romantically. It contained waste, fraud, poor designs, political favoritism and weapons that arrived late or did not work. The relevant comparison is not between a flawless wartime system and a corrupt modern one. It is between the outputs and institutional structures of two imperfect systems.
The wartime system absorbed inefficiency while expanding production at extraordinary speed. It converted civilian plants, financed new capacity, standardized designs, multiplied sources of supply and treated industrial bottlenecks as national problems rather than as the proprietary concerns of a single prime contractor. The War Production Board allocated scarce materials and coordinated priorities across the economy. The Reconstruction Finance Corporation and its Defense Plant Corporation financed capacity that could be government-owned and contractor-operated. The armed services specified what they needed and accepted the weapons, but they did not alone control every function of industrial mobilization.
That distinction matters more than nostalgia about assembly lines. Washington used public capital to create public leverage. It did not assume that awarding a contract to an incumbent was the same thing as creating national productive capacity.
The contemporary system often does the reverse. It can absorb years of development cost, fund facilities, train workers, rescue suppliers and pay cost-to-complete bills without obtaining another producer, transferable technical data, a government-controlled production process or a practical ability to shift work. World War II procurement absorbed waste while expanding output. Modern procurement too often absorbs cost growth while accumulating backlog.
From an Arsenal to a Collection of Programs
The post-Cold War consolidation of the defense industry transformed the government's position. The Defense Department itself reported in 2022 that the number of major aerospace and defense prime contractors had fallen from 51 in the 1990s to five. Consolidation was defended as a rational response to lower demand, and in some sectors it probably was. But it also changed the nature of competition. The government increasingly holds competitions for the right to become a long-term monopoly supplier.
The Joint Strike Fighter illustrates the structure. The development competition was intense, but the prize was not an ordinary production order that would be re-competed periodically. It was control of a weapons ecosystem expected to endure for decades. The losing airframe did not create a continuing second source. Large firms that did not lead the winning bid could remain important suppliers, spreading work and preserving political support without preserving meaningful competition at the level of the complete aircraft.
There are legitimate reasons to subcontract. A prime must integrate thousands of components, enforce configuration control and accept responsibility for performance. The Federal Acquisition Regulation already states that the government will not pay “excessive pass-through charges” when a contractor adds negligible value, and it requires reporting when subcontracted effort exceeds 70 percent of contract cost in covered circumstances. Yet the larger problem is structural rather than a question of one fee. A competition is not necessarily healthy merely because several corporate logos receive revenue after award. If only one organization controls the design, interfaces, data and final integration, subcontracting can distribute income while leaving the government with a single practical source.
Nor is all ostensibly private research independent of public money. Federal rules permit reasonable and allocable independent research-and-development and bid-and-proposal costs to be charged as indirect expenses on government contracts. That is not inherently improper; the government benefits when established firms invest between programs. It does mean, however, that the contrast between taxpayer-funded incumbents and entirely self-funded newcomers should not be drawn too cleanly. The more useful distinction is who bears the consequence of delay and failure. A company that spends investor capital to build a product before a major order faces a discipline different from a contractor whose development workforce, overhead and schedule extensions remain supportable through government business.
The New Entrants—and Their Limit
The new defense firms are changing that risk allocation. Anduril raised $1.5 billion in private capital in 2024, with much of it intended to support a large manufacturing facility for autonomous systems. In the Air Force's Collaborative Combat Aircraft competition, Anduril and General Atomics advanced while three traditional primes did not; companies wishing to continue outside the funded phase would have to use their own resources. The significance is not that every new entrant is automatically efficient. It is that private capital, iterative product development and production engineering are again becoming part of defense competition.
The same pattern is appearing at sea. New shipbuilders are pursuing autonomous vessels and new yards using commercial production methods. These efforts can broaden the maritime industrial base, but a company capable of producing an unmanned surface vessel cannot quickly become a second builder of nuclear submarines. Nuclear certification, specialized facilities, a trained workforce, naval reactors, design data and thousands of qualified suppliers form barriers that venture capital cannot cross on a normal investment horizon.
Traditional primes therefore have several rational responses to the new competition. They can acquire successful newcomers; become their indispensable integrators; control the interfaces through which new products connect to legacy platforms; or concentrate on programs whose capital requirements and regulatory barriers protect them from entry. None requires a conspiracy. Each follows from incentives.
The danger is a two-tier arsenal. Startups contest the fast-moving, comparatively inexpensive layer. Incumbents retain the slow-moving platforms without which the smaller systems cannot deploy, persist or survive. The Pentagon may receive better drones while remaining unable to obtain submarines on time.
The Evidence at the Industrial Core
Naval construction shows what captivity looks like in practice. In April 2026, the Congressional Budget Office reported that the Navy and industry had recently possessed the resources to build ships but had been unable to deliver them on time. The Navy's 2025 plan contained more than $10 billion in “cost-to-complete” funding for ships that Congress had previously authorized and fully funded. After inflation, CBO found price increases across nearly every major shipbuilding program it examined.
Exhibit 1 — Real increase in price per ship
Construction times tell the same story. Destroyers and submarines that took five to six years to build in the 2000s now take nine to ten. Ford-class carriers are taking ten to eleven years, compared with seven to eight for their Nimitz-class predecessors. CBO estimated that if ships were still completed at historical rates, the fleet would already exceed 300 ships; between 2026 and 2030, slower construction produces an average shortfall of 20 ships, principally destroyers and submarines.
The submarine figures are especially revealing. About 70 percent of suppliers for critical submarine components have no competitor. Attack-submarine production averaged only 1.1 boats a year over the preceding two years even though the Navy had generally been buying two per year. The service plans roughly $10 billion in additional support for the new-construction industrial base and another $8 billion for maintenance. More than 20 companies are now participating in outsourced submarine work, an encouraging form of distributed production, but the final system still depends on two nuclear-capable yards owned by two corporations.
This is not proof that the shipbuilders alone caused every delay. The Navy changes requirements, begins construction with immature designs, creates erratic demand and struggles to manage programs. Congress inserts work, protects facilities and changes annual quantities. Labor shortages, supplier failures and the pandemic have all mattered. A serious reform argument must acknowledge that the government is often an undisciplined buyer.
But shared responsibility does not eliminate monopoly leverage. It intensifies it. When a program is late, the government cannot credibly walk away if walking away would create an even larger hole in the fleet. The contractor becomes more essential as its performance deteriorates, and public funds used to repair the industrial base can make the same supplier still more indispensable.
Exhibit 2 — The captive-customer cycle
This is the central paradox of modern defense industrial policy: poor output can become an argument for giving the existing structure more money because the country cannot tolerate the consequences of its collapse.
Why the Pentagon Cannot Police This by Itself
The Department of Defense must remain responsible for military requirements, operational testing, acceptance and the employment of forces. It should also possess first-rate contracting expertise. What it cannot do reliably is serve simultaneously as customer, market regulator, rescuer and sole evaluator of whether its own acquisition strategy worked.
The conflict is practical, not moral. A Navy official deciding whether a shipbuilder requires additional investment is not considering industrial structure in the abstract. That official is considering a submarine or carrier the fleet must have. Refusing support may strengthen negotiating discipline in theory while worsening a near-term military shortfall. The operational mission will usually—and understandably—win. The supplier knows this.
GAO's 2025 assessment found that the Defense Department planned nearly $2.4 trillion for 106 of its costliest weapon efforts, while the average time for major programs to provide even an initial capability had risen to almost twelve years. This is not simply a failure of individual contracting officers. It is the consequence of asking the buying institution to impose market discipline after strategic dependence has already formed.
Congress holds the purse strings, but ordinary congressional oversight is also insufficient. Members must reconcile national industrial policy with jobs, districts, committees and annual appropriations. GAO and CBO provide indispensable analysis, but neither possesses executive authority to condition or redirect industrial investment. The Federal Acquisition Regulatory Council writes government-wide rules; it does not decide whether the nation needs a second submarine-component source or should own a critical shipyard facility.
The missing institution lies between analysis and appropriation.
Put Industrial Leverage on the Purse-Strings Side
Congress should create a National Arsenal Board in the Executive Office of the President, supported by an independent Defense Industrial Scorekeeper serving Congress. The names matter less than the separation of functions.
The Board should not choose tactics, write military requirements or tell a service which weapon it needs. Its jurisdiction should begin when public money is requested to sustain, rescue or expand a structurally concentrated industrial base. It should certify whether the proposed expenditure increases national capacity, competition and government leverage—or merely capitalizes an incumbent.
For any major industrial-base subsidy, cost-to-complete appropriation, facility investment or extraordinary schedule relief, the Board should be required to answer several questions. Does the government receive transferable technical data or escrow rights? Can another qualified producer make the component? Does public money create a second source, a government-owned facility, surge rights or step-in authority? Is the production architecture modular enough to distribute work? What measurable output—ships delivered, modules completed, maintenance days reduced—will release the next tranche of funds?
The congressional scorekeeper would audit those certifications and publish a common set of program measures across the services. It would distinguish money obligated from capacity created, workforce hired from output delivered, and subcontract dollars distributed from genuine alternative sources established. Congress could then release industrial funding in tranches tied to certified milestones rather than folding subsidies into program accounts where performance is difficult to isolate.
This arrangement would restore a useful wartime distinction. DoD would specify, buy, test and operate. A government-wide body, accountable to the President and Congress, would manage the public's industrial leverage. Independent scorekeepers would judge results using measures that neither the service nor the contractor controls.
Exhibit 3 — A divided system of authority
Function | Responsible institution | Governing test |
Military need, requirements and operational test | DoD and the military services | Does the system meet the warfighter's need? |
Industrial investment, second sources and public facilities | National Arsenal Board | Does public money create capacity and leverage? |
Appropriations and conditional funding | Congress | Are funds released against measurable results? |
Independent cost, schedule and market assessment | CBO, GAO and a Defense Industrial Scorekeeper | Did the program deliver what was promised? |
Production and integration | Competing firms, arsenals and GOCO facilities | Can output scale, surge and transfer? |
The Rules That Should Follow
The new institutional structure would be meaningless without enforceable rules.
First, public rescue money must reduce public dependence. If taxpayers finance a plant, a production line or the recovery of a critical supplier, the government should obtain an ownership interest, step-in rights, technical data, surge capacity or a qualified alternative source. A subsidy that leaves the same bottleneck under the same exclusive control is not industrial policy. It is capitalization without leverage.
Second, the government should own or control the interfaces that make competition possible. Open architectures are valuable only when another firm has the data, tools, test environment and legal right to build against them. Data escrow should activate upon chronic default, insolvency, abandonment or a national emergency. For keystone platforms, the government must be able to replace subsystems without seeking the permission of the platform monopolist.
Third, design, integration, production and sustainment should not automatically become a single lifetime franchise. In some programs one firm must perform all four roles, but that outcome should require justification. The default should be periodic competition for upgrades and sustainment, government-controlled test infrastructure, and production packages designed for transfer.
Fourth, a second source must be real. Awarding subcontracts to additional firms does not create competition if those firms cannot deliver a complete module or component directly to the government. Congress should fund qualification, tooling and minimum sustaining orders for selected bottlenecks, particularly the critical submarine suppliers that currently have no competitor. Stable demand is essential: the government cannot demand that industry hire and invest while changing quantities every budget cycle.
Fifth, acquisitions of innovative defense firms by dominant primes should receive a national-security competition review that considers future markets, not only current revenue. A startup with modest sales may nevertheless be the only plausible challenger in an emerging category. Integration can sometimes accelerate fielding, but acquisition should not become the routine mechanism by which the old structure absorbs the new one.
Sixth, private risk should be rewarded, not merely celebrated. Companies that bring working prototypes developed substantially with private capital should receive faster test access, clear production criteria and a genuine path to scaled orders. At the same time, “commercial” branding should not excuse weak testing or hidden lock-in. The standard is not whether a firm is new; it is whether its model gives the government more speed, replaceability and leverage.
Seventh, allied ship purchases should be used as a bridge to domestic capacity, not as a substitute for it. Buying a proven icebreaker, auxiliary or small combatant from an ally can fill an urgent gap. The agreement should also transfer design knowledge, train American workers, establish domestic module production and qualify additional suppliers. Otherwise the United States merely exchanges a domestic bottleneck for an overseas dependency.
Finally, government must impose discipline on itself. Requirements should be stabilized before construction. Programs should avoid simultaneous invention, detailed design and production. Ship orders should be smoothed across years to retain skilled labor. A contractor should not be punished for idle capacity that the government's own boom-and-bust purchasing created, nor rewarded for delays caused by immature designs the government accepted. An honest allocation of responsibility is a prerequisite for an honest price.
Competition at the Edge Is Not Enough
The defense-tech movement can improve American military power. It can restore private initiative, shorten product cycles and challenge the assumption that every useful system requires a decade-long development program. Its success should be encouraged.
But a swarm of inexpensive autonomous systems does not repair a submarine in dry dock, build a nuclear reactor compartment or create a second source for a specialized valve. If the new firms flourish only in markets the traditional primes are willing to yield, the United States may celebrate a procurement revolution while the foundations of military power become more concentrated and fragile.
The answer is not to break up every large contractor or to recreate 1942. Nuclear ships and advanced aircraft require scale, integration and patient capital. The answer is to stop confusing corporate scale with national capacity and contract spending with output. Where competition is impossible in the near term, government must behave as the steward of a strategic utility: controlling interfaces, owning essential capacity where necessary, measuring output independently and ensuring that public investment buys public leverage.
The decisive reform is therefore constitutional as much as contractual. The institution that needs the weapon most urgently should not be the only institution deciding whether the supplier is efficient, whether a bailout is justified and whether the resulting monopoly is acceptable. Those judgments belong closer to the purse strings, supported by independent evidence and enforceable conditions.
DoD should be a demanding customer—one capable of stating a need, testing the result and walking away when a supplier fails. It cannot do that while it is also the captive customer, industry regulator, bailout authority and final judge of contractor performance. Unless the federal government separates those roles, innovation at the edge may simply make monopoly at the core easier to tolerate.
Sources and further reading
Congressional Budget Office, “Testimony on Challenges Facing the Navy’s and Coast Guard’s Shipbuilding Programs and the Shipbuilding Industrial Base,” April 22, 2026.
Congressional Budget Office, “An Analysis of the Navy’s 2025 Shipbuilding Plan,” January 6, 2025.
U.S. Government Accountability Office, “Weapon Systems Annual Assessment: DOD Leaders Should Ensure That Newer Programs Are Structured for Speed and Innovation,” June 2025.
Department of Defense, “State of Competition Within the Defense Industrial Base,” February 2022.
Federal Acquisition Regulation, “Limitations on Pass-Through Charges,” FAR 52.215-23.
Federal Acquisition Regulation, “Independent Research and Development and Bid and Proposal Costs,” FAR 31.205-18.
War Production Board, Wartime Production Achievements and the Reconversion Outlook, October 1945.
National Archives, “Records of the War Production Board.”
U.S. Government Accountability Office, “F-35 Sustainment: DOD Faces Several Uncertainties and Has Not Met Key Objectives,” April 2022.
Anduril Industries, “Anduril Raises $1.5 Billion to Rebuild the Arsenal of Democracy,” August 2024.
Reuters, “US Air Force narrows list to General Atomics and Anduril to build autonomous aircraft,” April 24, 2024.
About the author
The author writes on maritime security, strategy and industrial capacity. His work has appeared in the Center for International Maritime Security and the Middle East Naval Affairs Forum. He is the author of Blue Web Doctrine, available on Amazon.
Comments