There’s a food pantry on a base near you. There’s a food pantry on most of them. Operation Homefront, the Armed Services YMCA, the local Feeding America affiliate, sometimes the chaplain’s office running a closet out of a side room. The volunteers know which weeks are bad — paydays, PCS moves, deployment gaps that didn’t get paid out cleanly. They know which families come in for the first time and try to pretend they’re picking something up for a neighbor.
About one in four active-duty military households experiences food insecurity in a given year (USDA Economic Research Service comparison study using 2018 and 2020 data; RAND’s 2023 follow-up puts the figure at roughly 26 percent of active-duty service members). The civilian rate for a demographically comparable population is about one in ten. The gap isn’t an anomaly. It’s been measured the same way for years.
The United States spent $997 billion on its military in 2024 (Stockholm International Peace Research Institute, Trends in World Military Expenditure 2024). It is the highest figure SIPRI has ever recorded for any country in any year.
Both of those things are true at the same time. They are descriptions of the same system.
This is the same argument as the last two posts. The locks just got more expensive.
The Scale, Stated Plainly
The federal government does not have one number for what it spends on defense, which is the first useful thing to know. The FY2025 National Defense Authorization Act authorized $895 billion in defense discretionary funding, $921 billion when you add the mandatory funding swept up in the bill (Congressional Research Service, January 2025). SIPRI’s $997 billion total adds categories CBO doesn’t — military retirement, some international affairs lines — to capture what the country actually spends on its military function across the whole government. They aren’t contradictory; they describe overlapping circles. The number people argue about depends on which circle they’re in.
Pick the most conservative figure and the picture still looks like this. The United States spent more on its military in 2024 than the next nine countries combined (SIPRI). About 37 percent of all military spending in the world, by every country, was done by one country. That ratio is compressing — by SIPRI’s 2025 numbers the US outspends “only” the next six combined as Germany and Ukraine have ramped — but the absolute number keeps rising regardless of who else is spending what.
Military spending eats roughly 3.4 percent of GDP, higher than any other developed country not currently at war. It runs at about $2,895 per American per year. It accounts for more than half of all federal discretionary spending — the spending Congress actually votes on each year, as opposed to mandatory programs like Social Security and Medicare. Whatever else Congress fights about, it does not fight about this. The NDAA passes every year with bipartisan supermajorities.
The base budget is one part of the bill. The wars are the other. Brown University’s Costs of War Project puts the cumulative cost of the post-9/11 wars in Iraq, Afghanistan, Pakistan, Syria, and the smaller counterterrorism operations at roughly $8 trillion through fiscal 2022, not counting future interest on the borrowing that funded them. Care for the veterans of those wars is projected to add another $2.2 to $2.5 trillion through 2050. Interest on the war debt has already cleared a trillion dollars on its own. None of those are DoD line items in a given year. But they’re the bill, and it gets paid.
Where The Money Actually Goes
Three things absorb most of the spending. They overlap.
The first is the contractors. Between 2020 and 2024, private firms received $2.4 trillion in Pentagon contracts — roughly 54 percent of all defense discretionary spending in that period (Costs of War Project, Profits of War, 2024). Five firms took $771 billion of that. Lockheed Martin pulled in $313 billion across those five years. RTX (the former Raytheon) $145 billion. General Dynamics $116 billion. Boeing’s defense unit $115 billion. Northrop Grumman $81 billion. That’s an average of about $154 billion a year going to five companies — more than the entire annual military budget of any country in the world except China.
The second is the flagship procurement program. The F-35 is now projected to cost over $2 trillion across its full lifetime — $442 billion to acquire roughly 2,500 aircraft, plus $1.58 trillion to operate and sustain them through 2088 (Government Accountability Office, April 2024). The estimate keeps going up. In 2018 GAO put sustainment at $1.1 trillion. By 2023 it was $1.58 trillion, a 44 percent increase, partly because the Pentagon decided to fly the airplane eleven years longer than originally planned. The F-35A’s mission-capable rate — the fraction of the fleet ready for any assigned mission at a given moment — was 52 percent in 2023. The program’s own minimum target is 80 percent.
The third is the part of the system the books don’t seem to know about. The Department of Defense received a disclaimer of opinion on its consolidated financial audit in November 2024, the seventh year in a row since the agency was first required to undergo a full audit in 2018. A disclaimer of opinion means the auditors couldn’t gather enough evidence to form any opinion at all on whether the financial statements were accurate. The DoD comptroller has argued publicly that “failed” is unfair, since 9 of the 28 sub-entities did receive clean opinions and the agency claims progress toward a clean overall audit by 2028. The 15 entities that received disclaimers cover a system with $4.1 trillion in assets and $4.3 trillion in liabilities. Nine of twenty-eight passed. Pick your preferred verb.
How It Stays Big
The political mechanism is straightforward to describe, which is part of why it’s so durable. None of it is hidden.
Major weapons systems are built with subcontractor networks distributed across as many states and congressional districts as possible. The F-35 supply chain runs through 45 states and somewhere north of 1,650 suppliers by Lockheed Martin’s own count, supporting roughly 290,000 jobs. That isn’t a coincidence of where aerospace engineers happen to live. It’s a deliberate design. Voting to cut the program means voting to cut jobs in your district, and in every other member’s. The program survives because cutting it costs every member of Congress something concrete.
The contractor money runs through the standard channels and also through the less standard ones. Lobbying is the most visible piece. The arms industry employed about 950 registered lobbyists in 2024 (Costs of War Project, tracking Senate lobbying disclosures). Campaign contributions go to members of both parties, weighted toward the committees that matter — Armed Services, Defense Appropriations. And then there’s the revolving door: senior Pentagon acquisitions officials and retired flag officers moving into contractor board seats, defense-focused think tank fellowships, and lobbying roles within a couple of years of leaving government. The pattern isn’t disputed. It’s the org chart.
The annual NDAA — the bill that authorizes all of this — passed the House in December 2024 by 281 to 140 and the Senate by 85 to 14. The year before that, 310 to 118 in the House. The year before that, in 2022, it cleared the House 350 to 80. Whatever else Congress fights about, year after year, this comes through.
What It Isn’t Buying
Now go back to the base, and the food pantry.
Junior enlisted pay starts around $24,000 in base salary for a private in year one, climbing to roughly $34,000 for an E-4 at four years of service. There are housing and food allowances on top of that, but the housing allowance varies enormously by base location and the food allowance is calibrated for a single service member, not a service member with a spouse and two kids. The Basic Needs Allowance, created in 2022 to bridge the gap for the lowest-paid families, excludes the housing allowance from the income calculation for some federal purposes but counts it for others. Eligibility is narrow. One in four military households remains food-insecure, with the rate highest among the junior enlisted ranks.
The Department of Veterans Affairs runs the country’s largest integrated healthcare system. The disability claims backlog — the queue of cases waiting more than 125 days for an initial decision — fluctuates in the high tens to low hundreds of thousands depending on how it’s counted, and has been a chronic problem across administrations of both parties. The post-9/11 veteran suicide rate runs about one and a half times the rate for demographically comparable civilians, by the VA’s own annual suicide-prevention reports.
The services have missed their recruiting targets for several years running. The Army, the largest service, has had multiple consecutive years of shortfalls, and the response has been higher enlistment bonuses, lower educational requirements, and a public-relations campaign. Pay reform has been on the table. The FY2025 NDAA’s pay raise for junior enlisted servicemembers — 14.5 percent for E-4 and below — was the first meaningful adjustment in years, and it isn’t clear yet whether it’ll move the recruiting needle. It’s a long way from “the troops are taken care of.”
A spending number that grows every year. Lockheed’s operating margin is one thing. A junior enlisted family in the pantry line at the base town is another. They’re descriptions of the same budget.
What’s Real
None of this is an argument that the United States doesn’t need a military. It does. Treaty commitments to NATO and Pacific allies are real obligations with real deterrent value, and credible commitment is part of how wars get prevented rather than fought. Some weapons systems work as designed and represent real engineering achievements. A great many service members and officers are doing work that is genuinely important and that no one else is going to do.
It’s also not an argument that defense contractors are bad people who shouldn’t exist. Building advanced military hardware is hard, capital-intensive work, and a domestic industrial base for it has strategic value beyond any single program. Some of the cost overruns on programs like the F-35 are caused by changing requirements, technical challenges, and the difficulty of integrating systems no one has built before. Not all of the inefficiency is fraud. Some of it is the cost of doing the actual thing.
The argument is about scale, structure, and direction. A budget of $997 billion isn’t the same conversation as a budget of $600 billion. A cost-plus contract isn’t the same instrument as a fixed-price one. A weapons program built with subcontractors in 45 states isn’t the same political object as one built in three. None of the qualifications above changes the picture at the scale we’re operating at.
What They’re Paying For
The product that nearly a trillion dollars in annual military spending is purchasing is not, primarily, defense. The country was already militarily unmatched at half this number. The purchase is three things stacked together.
It’s a permanent industrial subsidy to a small number of large defense contractors who function as quasi-utilities — guaranteed customers, guaranteed margins, guaranteed political protection. They aren’t competitive enterprises in any meaningful sense. They sell to one customer, and that customer has been instructed by Congress, every year, to buy.
It’s a system of political patronage that distributes jobs across nearly every congressional district in a way that makes the spending self-sustaining. The jobs are real. The communities depending on them are real. But the system is structured to make the spending impossible to cut, not to make the spending efficient. Those are different design goals, and the one that won is the one that survives the budget process.
And it’s global force projection — 750-plus overseas bases in 80-plus countries, by the Costs of War Project’s running count — that is no longer tightly coupled to any single articulable strategic doctrine. Some of those bases serve real deterrence purposes. Some are there because they have been there for seventy years and no one has the political incentive to close them. Both kinds get the same budget line.
None of this is what’s on the box. The box says “national defense.” The contents are a subsidy, a patronage system, and a global posture. The actual defense — the thing that keeps Americans safe — is a smaller line item inside the bigger one.
The Fixes Are Boring
The fixes below are going to sound impossible. They sound impossible because the people who profit from the current arrangement have spent decades making them sound impossible — not because they’re radical, untested, or beyond the capabilities of any functional government. Every one of them is being done somewhere, has been done at some point in American history, or is the obvious structural answer that any serious accounting would arrive at. Smaller, politer versions of each have been tried already. They didn’t fix the lock. They were never designed to.
- Cut the defense budget in half over five years. Cap it at roughly 1.5x what the next-largest spender (currently China) puts into its military. The United States was militarily unmatched at $500 billion in real dollars. It is militarily unmatched at $997 billion. The difference between those two numbers is the subsidy, not the defense. “Audit and find some savings” has been the official position for thirty years; the budget has roughly doubled in that period.
- End cost-plus contracting. Not “for mature programs.” Entirely. Fixed-price competitive bids, with real financial penalties for overruns and delivery failures, and the cancellation of programs that miss them. The argument that complex programs need cost-plus is the same argument every regulated industry makes for the structure that guarantees its profits.
- Close most of the overseas bases. Maintain a small set in actual strategic partners, by mutual agreement, with defined missions. Close the rest. The 750-base global posture isn’t load-bearing defense infrastructure — it’s an inheritance from a Cold War that ended in 1991 and a counterterrorism doctrine that no longer matches the threat landscape. Every base has a host-nation cost, a political cost, and an opportunity cost.
- Pay enlisted personnel a living wage. A junior enlisted family in 2026 should not need food stamps and a base-town pantry while contractor CEOs clear eight figures. Set the floor at a real living wage for a household of four in the cheapest base location in the country, and let everything scale up from there. The current Basic Needs Allowance, with its housing-allowance carve-outs and narrow eligibility, is the cosmetic version. Replace it with the real one.
- Ban the revolving door with teeth. Ten-year prohibition on senior Pentagon civilians and flag officers transitioning into defense-contractor employment, board seats, lobbying, or paid advisory work. The current one-to-two-year cooling-off period is so trivial that nearly every senior official walks straight through it. Ten years means people retire from one career and pursue a different one, the way the rest of the workforce does.
- Require a clean Pentagon audit before appropriations. Not “encourage” auditing. Not “set a goal for 2028.” Pass a clean consolidated audit, and the next year’s funding flows. Don’t pass, and the funding is withheld until the audit closes. The 1990 Chief Financial Officers Act has been on the books for thirty-five years. Every other federal agency has figured this out. The only thing keeping the Pentagon from doing the same is that no one has actually withheld the money.
- Sunset every AUMF currently in force. The 2001 Authorization for Use of Military Force was written in the week after 9/11 to authorize the response to 9/11. It is still being cited to justify operations in countries that did not exist as policy concerns at the time. Sunset it. Sunset the 2002 Iraq AUMF. Require Congress to vote on every new use of force, with a defined geography, a defined mission, and a defined end date.
- Nationalize the top five primes, or break them up. The largest defense contractors are not commercial enterprises in any meaningful sense. They have one customer, guaranteed margins, and political protection built into their supply chains. Either bring them inside government, the way most peer countries handle their defense industrial base, or break them up into smaller firms that have to actually compete for fixed-price work. The current arrangement is a privatized utility with all the inefficiency of the public version and all the executive compensation of the private version, and the taxpayer is paying for both.
Who Is This For
The next post hasn’t been picked yet. The frame doesn’t change. Find the industry, ask what it’s actually delivering at the scale of money flowing through it, and ask the question this series has been asking since Part 5: who is this for. The food pantry on the base is one answer. The $25 million executive comp package is another. They share a budget line. They just don’t share a beneficiary.
Even that's Odd 
Leave a comment