$90 Billion a Year, Managed by Nobody

Bottom line up front. We spend roughly $90 billion a year maintaining weapon systems, and almost no program has ever written down who is supposed to be doing that work in year 20. This is not for lack of policy. Statute and instruction already require lifecycle product support planning, and have for years. What they require is a set of activities. What no document requires is a named performer, on a dated handoff, with the enabling purchase funded in the budget we are building now. That gap is why the depot cannot fix a tank without the manufacturer. The fix is one artifact: a sustainment transition plan, written before Milestone B, that names who performs which work across four phases, prices technical data as options against those named handoffs, uses early field data to target which parts actually need a second source, and times partner demand to keep those suppliers alive. One rule governs it. Buy one phase early.

A government depot can strip an M1 Abrams to the bare hull and build it back. Sometimes it cannot fix one.

The Government Accountability Office (GAO) reviewed 18 Army and Marine Corps ground vehicles in 2026 and found that all 18 had trouble getting parts and all 18 had trouble getting technical data (GAO-26-108888). Depot maintainers were sending some Abrams, Bradley, and Stryker work back to the manufacturer. Not for lack of skill, tooling, or floor space; the drawings were proprietary.

The Littoral Combat Ship shows the same gap earlier in the arc. GAO found the Navy had not determined which maintenance tasks its own sailors would perform versus contractors, that operating and support estimates did not account for the revised maintenance approach, and that actual cost data were incomplete (GAO-22-105387). The support concept changed and no document said who does what, when.

Neither is a maintenance failure. Both are the predictable result of never deciding, at any point in the program, who was supposed to be doing that work in year 20.

The Department of War spends nearly $90 billion annually on weapon-system maintenance (GAO-23-105556). We manage almost none of it against a plan that names anyone.

This is not a new idea. That is the problem.

Anyone who has worked sustainment will read the preceding paragraphs and think, correctly, that we already have policy for this. We do. That is exactly what makes the failure worth examining rather than simply restating.

Title 10 U.S.C. § 4324 requires a product support strategy, performance goals, a lifecycle cost estimate, engineering and design considerations, an intellectual property (IP) management plan, and a strategy for using public and private capabilities. DoD Instruction 5000.91 goes further, requiring product support planning before program initiation and establishing the Life Cycle Sustainment Plan (LCSP) as the primary program management reference for operating and support planning and execution from program inception to disposal. The LCSP is already supposed to carry sustainment metrics, risks, costs, affordability constraints, obsolescence and supply chain risk, engineering and design considerations, technical data and IP, and maintenance and overhaul requirements.

On paper, the problem is solved. In the field, GAO keeps writing the same report.

The Army did not acquire the rights to compete Stryker spare parts at award, and when it later sought alternatives to contractor support the contractor declined to sell the package and the program assessed that buying it had become prohibitively expensive (GAO-22-104752). The F-35 program lacked a comprehensive understanding of which technical data rights it owned, which it still needed, or what they would cost. Navy officials said they did not fully understand what IP they needed until systems broke and maintainers could not repair them (GAO-25-107468).

Every one of those programs had an LCSP.

Nor is the pattern confined to individual programs. The Department still has limited visibility into where goods are manufactured or whether lower-tier suppliers are domestic (GAO-25-107283), and lacks a department-wide approach to tracking whether identified supplier risks were ever mitigated (GAO-22-104154).

So the honest question is not what additional policy to write. It is why the policy we have does not bind. Four reasons, and they compound.

The documents require activities, not performers. An LCSP describes a product support strategy. It does not have to say that in year 12 the depot at a named location performs this repair, or that a qualified vendor other than the prime supplies this component. Activities can be documented indefinitely without anyone ever being on the hook.

The purchase and the payoff sit in different budgets. The technical data option that enables a Phase 2 handoff has to be bought during Phase 1, which means it competes against unit cost and schedule in a program that is being evaluated on unit cost and schedule. The benefit lands eight years later. No one currently has to reconcile those two facts in the same document.

The incentive horizon is shorter than the lifecycle. This is the part policy cannot fix by itself. The program manager who would fund the data option will not be in the seat when the handoff arrives. Neither will the program executive officer, the contracting officer, or the requirements sponsor. A forty-year commitment is being made by people whose accountability window is measured in single-digit years, and nothing in the current artifact set survives their departure with enough specificity to be enforced.

Someone is always better off without the handoff. Role transition moves work, and whoever performs that work today rarely welcomes the change. Some of that resistance is legitimate. Incumbents raise real arguments about safety, configuration control, qualification, and liability, and they are sometimes right. Some of it is simply that recurring revenue is being reallocated, and the party losing it has more staff, more continuity, and more time than the officer proposing the change. Either way, a plan that assumes goodwill will lose to a plan that anticipates opposition.

Those are the failure mechanisms. The first three are structural rather than deliberate. The fourth is deliberate, and any proposal that does not plan for it was written by someone who has never tried to move the work. What none of them is, is a gap in authority. We have never made the handoff itself a named, dated, funded object that outlives the people who created it and survives the people who would rather it did not happen.

There is one place we already do this, and it proves the concept works. DoD Instruction 5000.91 requires that software product support include early planning for data rights, source code, and technical data, and specifically for the ability to transition to another developer when applicable. That is a transition plan. We wrote the requirement, for software, because the vendor-lock problem became impossible to ignore. The proposal here is to extend a requirement the Department has already accepted in principle to the rest of the system.

What General Motors plans that we do not

A new vehicle enters service on a schedule its manufacturer already understands.

For the first three years the dealer network handles nearly everything, and every warranty claim feeds back to reliability engineers who redesign the weak components. Around year three the independent shops, parts remanufacturers, scan-tool vendors, and aftermarket suppliers begin absorbing the work. By year eight most repairs happen outside the dealership. Past year ten the original equipment manufacturer (OEM) stops trying to stock every component and the remanufactured-parts and salvage market carries the fleet. S&P Global Mobility puts 289 million light vehicles on U.S. roads at an average age of 12.8 years, which is the demand pool making all of it economic.

None of that is charity or accident. General Motors designs vehicles around components other firms can build, publishes service information, produces diagnostic procedures and technical data packages, and sets maintenance intervals, all before the first unit sells. The handoff is engineered during development. The market then executes a plan that already exists.

It also does not run on goodwill. The Auto Care Association found that 84 percent of independent repair shops rank data access as the top issue in their business, and that 51 percent send as many as five vehicles a month back to dealers because of data restrictions. That is the identical failure to a depot shipping an Abrams back to its manufacturer, and the commercial world only partly solved it through Environmental Protection Agency service-information rules, a negotiated right-to-repair memorandum, and continuing argument that the Federal Trade Commission documented in Nixing the Fix. Access was engineered, litigated, and regulated into existence.

So the transferable lesson is not "open the aftermarket." It is that a sustainment ecosystem never emerges from ownership alone. It emerges when scale, demand, technical information, tooling, parts access, qualification, and incentives make third-party support viable, and somebody decides during design that it will.

We do the first phase and skip the rest. For the Assault Amphibious Vehicle (AAV), the Light Armored Vehicle (LAV), the Amphibious Combat Vehicle (ACV), and most of the fleet, the program office keeps carrying sustainment indefinitely. Some of that is unavoidable. Our fleets are small, our demand is thin, our configurations are controlled, and security limits who can touch the system. But a large share of it is simply that nobody wrote down who takes the work next, so nobody bought what the next performer would need.

Scale is the part we cannot copy. The plan is the part we can.

The sustainment transition plan

One artifact, built before Milestone B, updated at every sustainment review. It states who performs which sustainment work at each phase of the system's life, and what must be acquired in each phase to make the next one possible.

Phase 1. OEM-dominated support, fielding to roughly year five. The prime does nearly everything, deliberately. The system is immature, failure modes are unknown, configuration is still moving. But the purpose of this phase is not support. It is producing the target list: which of the many thousands of parts on the system actually drive readiness, and therefore which ones are worth buying data and qualifying sources for.

We are not producing that list. GAO found that as of June 2022 predictive maintenance had not been fully implemented on any weapon system it reviewed, that the services had not consistently tracked adoption, and that they generally lacked metrics to evaluate results (GAO-23-105556). Where it has been done it works: the Air Force Rapid Sustainment Office reports using artificial intelligence (AI) to improve maintenance data quality, evaluate sensor and maintenance history, predict component failures, and save thousands of maintenance hours a year. The Department's own VAULTIS standard already sets the test, requiring data that is visible, accessible, understandable, linked, trustworthy, interoperable, and secure. Phase 1 fails quietly when the fleet is not instrumented at fielding, because the program reaches year five with no evidence about what to buy.

Phase 2. Shared support, roughly years five through fifteen. This is the transition window and the phase we currently skip entirely. Field data has now identified the few hundred items consuming the readiness budget. Exercise the data-rights options on those items, not on the whole system. Qualify second sources and organic depot capability against them, with test procedures and acceptance criteria written down. Stand up public-private partnerships under 10 U.S.C. § 2474. Begin aligning partner demand to production windows while the line is still warm.

Architecture decides how much of this is possible at all. The Modular Open Systems Approach (MOSA) Implementation Guidebook ties modular open systems to competition, technology refresh, and interoperability, and severable modules with published interfaces are what let a program replace a subsystem without redesigning a platform. But open interfaces without qualification, configuration control, and cybersecurity produce risk instead of resilience. MOSA creates the possibility of competition in Phase 2. Data rights and qualification create the competition itself.

Phase 3. Distributed support, roughly year fifteen to replacement. Organic depots, qualified small businesses, and sustainment suppliers carry mature repairables, commodity parts, remanufacture, and obsolescence work. The prime holds what only a prime can hold: design authority, integration, safety-critical engineering, major modernization, and the replacement program.

The authorities are already written. 10 U.S.C. § 2464 states that a government-owned, government-operated core logistics capability is essential to national defense. Section 2474 authorizes the public-private partnerships that make it work in practice. Remanufacture belongs here too, and it is a mature industrial process rather than a hopeful one: the U.S. International Trade Commission defines it as restoring end-of-life goods to original working condition, and Caterpillar and Cummins run it at scale to recover material, lower customer cost, and fold in design improvements. Engines, transmissions, power distribution modules, sights, and line-replaceable units are candidates the moment the data, test procedures, and configuration control exist. The F-16 Service Life Extension Program is the defense version at scale, putting structural modifications into roughly 300 F-16C/D aircraft to extend them from 8,000 to 12,000 flight hours and keep them viable toward 2050. That is engineered lifecycle management, and it produced stable depot workload, tooling, and a trained workforce as a byproduct.

Phase 4. Legacy and partner support, replacement onward. The system remains valuable to allies for another fifteen or twenty years. Partner demand, remanufacture, and modernization packages keep the suppliers and depot lines that Phase 2 built from going cold.

The F-16 is the clearest proof this phase is real. After Fort Worth closed, partner demand supported a new Block 70/72 line in Greenville, South Carolina, at the time the only F-16 production facility in the world, building 128 aircraft for five partner nations under roughly $14 billion in awards, with 25 nations operating the type and 405 partner aircraft slated for V-configuration upgrades. The C-130J shows the same effect in the sustainment tail rather than the production line: its Selected Acquisition Report described 25 active foreign military sales (FMS) production and sustainment cases, with the program office managing more than $4.5 billion of work for 13 countries. Partner demand there is not a sale. It is multi-decade workload, and the Security Assistance Management Manual's trust fund structure is how it reimburses the activities performing it.

The Abrams supplies the caution. Department of Defense Inspector General work as far back as 1992 recorded the Army considering a combination of foreign sales and M1 remanufacture to keep production lines active between domestic buys, and warned in the same report that foreign sales alone would not sustain an active industrial base. Phase 4 is a contributor to industrial health, not a substitute for a requirement.

The phases are not the innovation. Every program manager recognizes them. The innovation is writing the handoffs down in advance and holding the program to a single rule.

The rule: buy one phase early

Every handoff in the plan requires something acquired in the phase before it. Miss the purchase and the handoff does not happen, regardless of how much anyone wants it to.

You cannot qualify a second source in Phase 2 if you did not price the technical data package in Phase 1. You cannot target the data buy in Phase 1 if the fleet was not instrumented at fielding. You cannot move depot work in Phase 3 if nobody wrote qualification criteria and test procedures in Phase 2. You cannot align partner demand in Phase 4 if the production window closed in Phase 3.

This reframes the data-rights argument, which we usually lose because we argue it as ownership. We do not need to own everything and cannot afford to. Data rights are options, priced against a named handoff on a known date. The question is never "should we own this?" It is "which handoff in the plan fails without it, and what will that option cost once we have no leverage?"

Leverage is highest before award and approaches zero afterward. The transition plan is what tells you, while you still have leverage, exactly which options to buy.

What makes this different from the LCSP we already write

If the answer to a failed document is another document, this proposal deserves the skepticism it will get. So the transition plan has to differ from its predecessors in four specific, checkable ways.

It names performers, not activities. Not "organic depot capability will be considered." Instead: this component, at this facility or this class of qualified vendor, beginning in this phase. A name can be wrong and later corrected. An activity cannot be wrong, which is why it never changes anything.

It attaches dates to handoffs. Each transition has a phase boundary and an approximate year tied to the program's own fielding schedule. Dates create the forcing function that "throughout the lifecycle" never does.

It converts enabling purchases into priced, budgeted options. Every named handoff carries the specific acquisition that makes it possible, with a cost, in the year it must be bought. This is the provision that puts a Phase 1 program manager on the hook for a Phase 2 outcome, and it is the one most likely to be resisted, because it is the only one that costs money now.

It fails at a gate. The plan is reviewed where reviews already occur, and a program that has not executed the enabling purchases from the prior phase is reported as not having executed them. Not a new board. An existing gate that produces a consequence.

None of that requires new authority. It requires the artifacts we already produce to name names, carry dates, price the options, and be checked. The Life Cycle Sustainment Plan is the natural home because DoD Instruction 5000.91 already designates it the primary program management reference from inception to disposal. It is simply not currently written in a form anyone can be held to.

Why this is not an attack on industry

A prime contractor's cost structure is built to develop capability. It carries systems engineers, integration laboratories, configuration and safety authorities, and the overhead that makes all of it auditable. That is exactly right for designing a new amphibious vehicle or certifying a structural modification.

It is a poor instrument for reproducing a cast bracket designed in 1994.

When we leave mature legacy parts on the prime by default, we pay next-generation overhead rates for commodity work. That is the visible cost. The invisible one is worse: the prime's best engineers spend their days closing legacy sustainment tickets on a platform we intend to replace, instead of building the replacement we are simultaneously asking them to build.

Both sides lose. Phase 3 is not a divorce. It is a division of labor that lets the prime do what it is actually built for. It is also why the plan should be negotiated at award, when the prime can see its own role across all four phases and price it, rather than litigated in year 18 when the only available move is to say no.

Where plans break

Two places, and both are failures of measurement rather than intent.

The first is confusing suppliers with capacity. The Department relies on a network of more than 200,000 companies, most of which it cannot see past the first tier. GAO has separately found that its data may not fully capture small-business participation through subcontracts and alternative agreements (GAO-22-104621). A firm can win an award and still lack the tooling, capital, cyber posture, or material access to build at volume. Counting suppliers measures the bottom rung of a ladder running identified, qualified, recurring, capitalized, secure, and surge-capable. Only the top rung is capacity. A Phase 2 plan that qualifies a supplier without giving it recurring demand has produced potential, not capability.

The second is treating partner interest as partner demand. General interest warrants monitoring. A signed and funded Letter of Offer and Acceptance is workload. GAO has documented both directions of the timing problem, finding that domestic and foreign requirements combined on one contract can lower prices and speed delivery, while a partner requirement that misses the production window reduces economies of scale and raises unit costs for everyone (GAO-17-682). Phase 4 only works when the demand is real, funded, timed, and close enough to our configuration to share a supply base.

What skipping the plan costs

Operating and support (O&S) costs run roughly 70 percent of a system's lifecycle cost, and GAO found critical O&S growth in 14 of the 36 systems it reviewed for fiscal years 2023 and 2024 (GAO-26-108140). The Congressional Budget Office has documented that aircraft grow more expensive to operate as they age, at rates that vary by fleet. Every trade made at Milestone B, data rights for unit price, proprietary interfaces for schedule, contractor support for a smaller organic footprint, is a loan drawn against that account at a rate nobody computes. Call it readiness debt. It is repaid thirty years later by a maintenance chief who never met the person who signed it.

Bridging without a plan carries its own bill. The AAV survivability upgrade was framed as a bridge to the future amphibious vehicle. The Inspector General later found it achieved its primary force-protection requirement but not all program requirements before initial production, including cost control, tactical egress, and reliability, and that legacy parts failures degraded reliability in ways that could reduce operational vehicles and increase spares demand and maintenance time (DODIG-2018-060). A bridge platform is only useful if the bridge holds, and holding it is a Phase 3 problem that has to be planned in Phase 2.

The debt is hard to see because we are bad at pricing the alternative. GAO compared two estimates for shutting down Bradley production and restarting it later. An Army study put it at $53 million. An earlier study by the manufacturer put it at $750 million.

That spread is not a rounding problem. It is assumptions about tooling, supplier requalification, workforce reconstitution, and restart schedule doing more work than the analysis. RAND found the same sensitivity examining the end of F-22A production, where shutdown, warm production, and restart options carried materially different profiles and restart carried substantial nonrecurring cost. A transition plan forces those assumptions into the open in Phase 2, when they are still cheap to test. Without a baseline, a savings claim is advocacy.

Why now

The Department is already asking for most of this, which is the strongest argument that the timing is right.

The 2025 Acquisition Transformation Strategy calls for expanding the industrial base, stabilizing demand signals, increasing commercial preference, engaging suppliers below the prime, reforming foreign military sales, using open systems architecture, mapping supply chains, improving government repair and technical data, and strengthening Life Cycle Sustainment Plans. It explicitly connects LCSPs to designing, building, and buying systems that are cost-effectively maintainable throughout the lifecycle. The National Defense Industrial Strategy frames the industrial base as a strategic ecosystem built on resilient supply chains, workforce readiness, flexible acquisition, and economic deterrence, and calls for considering exportability during design, which is Phase 4 planning by another name.

Read as a list, those are nine separate initiatives owned by different offices on different timelines. Read as a sequence, they are the transition plan. The strategy already asks for the pieces. It does not yet ask any single program to put them on one page, against named performers and dates, and be held to it.

Pilot it on five programs

Pick five programs with visible lifecycle exposure and make each one write the transition plan. Include a long-lived aircraft, a combat vehicle, a surface ship, a munition with a thin supplier base, and a partner-heavy platform, because phase timing differs across them and that variance is what we need to learn.

Each plan states four things. Who performs which sustainment work in each of the four phases, by name or by qualified class. Which technical data options get priced now, against which specific handoff, in which budget year. Which fielded-data streams will produce the Phase 2 target list, and who owns them. Which partner demand is credible enough to plan Phase 4 around.

Build it inside artifacts we already owe. 10 U.S.C. § 4324 already requires a product support strategy, a lifecycle cost estimate, engineering and design considerations, and an intellectual property management plan. DoD Instruction 5000.91 already makes the LCSP the governing reference. This is not a new document. It is those documents forced to agree with each other on dates and names, and then checked at a gate that already exists.

Then execute the Phase 1 actions and judge the pilot on one question. Did it change a decision? Which data option got priced. Which supplier got qualified. Which depot capability got funded. Which mature task moved. If the plans produced findings that changed nothing, the model failed. Say so and stop.

That test is the guardrail against the fair objection to all of this, which is that it is more process applied to a system that already moves too slowly. The answer is not a promise of restraint. It is a test the model can fail.

Nobody owns year 20

We are moving faster on procurement, and we should. Speed is a real advantage. But speed does not eliminate sustainment cost. It defers it into an account that is harder to see, harder to attribute, and paid by people who were not in the room and cannot renegotiate the terms.

Nothing here is blocked. Every instrument the plan needs is already written, staffed, and funded. They have simply never been made to agree with one another on a schedule, against named performers, before the decisions harden.

That is the finding. We have been writing sustainment policy against this problem for decades and the reports have not improved, because we kept producing better descriptions of activities instead of assigning the work to someone with a date attached. The document was never the missing piece. The name and the date were.

GM's aftermarket did not appear because the market is generous. It appeared because somebody decided during design who would be repairing the thing in year ten, and then bought what that person would need. We can decide the same thing. We have simply never written it down.

Everything above reduces to one artifact and one rule. Name who performs the work in each of four phases, before Milestone B. Then buy, in each phase, what the next one will require.

And it reduces further, to a single question that costs nothing and can be asked at the next program review.

When the briefer reaches the sustainment slide, ask who is performing this repair in year 20. Not what the strategy is. Who, by name or by qualified class. Then ask what we are buying this year to make that possible, and whose signature is on it.

If someone can answer, there is a plan.

If nobody can, you have just found another piece of the $90 billion.

About the author

Brandon Russell has spent twenty years in the United States Marine Corps in ground equipment maintenance, logistics, and readiness. He served in the Program Management Office for Light Armored Vehicles at the Army's Tank-automotive and Armaments Command, where he worked fleet sustainment, modernization, and condition-based maintenance on the legacy fleet and supported the development and acquisition of the Advanced Reconnaissance Vehicle, scheduled to begin replacing the LAV in 2031. He was the Marine Corps executive representative to the multinational LAV User Nations Group. As Battalion Maintenance Officer for 2d Light Armored Reconnaissance Battalion he was accountable for more than $120 million in equipment, directed over 4,200 service requests, and planned the accountability and validation of more than $200 million in assets. He designed and fielded a business-intelligence materiel readiness dashboard adopted across multiple units, and advised Force Design 2030 working groups on AI-enabled maintenance, condition-based maintenance, and data modernization. He was selected as the Marine Corps Ground Ordnance Maintenance Association Maintenance Officer of the Year for 2025. The views expressed here are his own and do not represent the Marine Corps or the Department of War.

Sources

Government Accountability Office

Army Combat Vehicles: Industrial Base Study's Approach Met Research Standards. GAO-15-548, 2015.

Foreign Military Sales: Expanding Use of Tools to Sufficiently Define Requirements Could Enable More Timely Acquisitions. GAO-17-682, 2017.

Defense Industrial Base: DOD Should Take Actions to Strengthen Its Risk Mitigation Approach. GAO-22-104154, 2022.

Small Business Contracting: Actions Needed to Implement and Monitor DOD's Small Business Strategy. GAO-22-104621, 2022.

Defense Acquisitions: DOD Should Take Additional Actions to Improve How It Approaches Intellectual Property. GAO-22-104752, 2022.

Littoral Combat Ship: Actions Needed to Address Significant Operational Challenges and Implement Planned Sustainment Approach. GAO-22-105387, 2022.

Military Readiness: Actions Needed to Further Implement Predictive Maintenance on Weapon Systems. GAO-23-105556, 2022.

Defense Industrial Base: Actions Needed to Address Risks Posed by Dependence on Foreign Suppliers. GAO-25-107283, 2025.

Weapon System Sustainment: DOD Can Improve Planning and Management of Data Rights. GAO-25-107468, 2025.

Weapon System Sustainment: DOD Identified Critical Cost Growth, and the Army Should Take Action to Yield Cost Savings. GAO-26-108140, 2026.

Military Readiness: DOD Should Take Further Actions to Address Challenges Across the Air, Sea, Ground, and Space Domains. GAO-26-108888, 2026.

Oversight and federally funded research

Department of Defense Inspector General. Review of the M1A2 Abrams Tank Program. Report 92-104, 1992.

Department of Defense Inspector General. Marine Corps Assault Amphibious Vehicle Survivability Upgrade. DODIG-2018-060, 2018.

RAND Corporation. Ending F-22A Production: Costs and Industrial Base Implications of Alternative Options. 2010.

RAND Corporation. Retaining F-22A Tooling: Options and Costs. 2011.

Congressional Budget Office. Operating Costs of Aging Air Force Aircraft. 2018.

Statute, policy, and strategy

10 U.S.C. § 2464, Core Logistics Capabilities.

10 U.S.C. § 2474, Centers of Industrial and Technical Excellence.

10 U.S.C. § 4324, Life-Cycle Management and Product Support.

Department of Defense. DOD Data Strategy. 2020.

Department of Defense. DoD Instruction 5000.91: Product Support Management for the Adaptive Acquisition Framework. 2021.

Department of Defense. National Defense Industrial Strategy. 2024.

Office of the Under Secretary of Defense for Research and Engineering. Implementing a Modular Open Systems Approach in Department of Defense Programs. 2025.

Office of the Under Secretary of War for Acquisition and Sustainment. Rebuilding the Arsenal of Freedom: Acquisition Transformation Strategy. 2025.

Service and agency reporting

Department of Defense. Selected Acquisition Report: C-130J Hercules Transport Aircraft. 2018.

Air Force Materiel Command. F-16 Service Life Extension Program a "Great Deal" for Department of Defense, Taxpayers. 2018.

Air Force Life Cycle Management Center. Air Force Opens New F-16 Production Line for Foreign Military Sales. 2021.

Rapid Sustainment Office. RSO Quarterly Report, April to June 2023. Air Force Life Cycle Management Center.

Defense Security Cooperation Agency. Security Assistance Management Manual, Chapter 9.

Commercial and regulatory

U.S. Environmental Protection Agency. Final Rule for Service Information Availability. 2003.

U.S. International Trade Commission. Remanufactured Goods: An Overview of the U.S. and Global Industries, Markets, and Trade. Publication 4356, 2012.

Federal Trade Commission. Nixing the Fix: An FTC Report to Congress on Repair Restrictions. 2021.

Auto Care Association. Survey: 84% of Independent Repair Shops View Vehicle Data Access as Top Issue for Their Business. 2024.

S&P Global Mobility. U.S. Vehicle Age Rises Again to 12.8 Years in 2025. 2025.

Caterpillar. Circular Economy: Remanufacturing.

Cummins. Cummins ReCon Remanufactured Engines and Parts.