Could the United States Really Run Out of Ammunition in a Month?

In the summer of 2024, the bipartisan Commission on the National Defense Strategy released a sobering assessment of America's defense capabilities. The report described war games involving a conflict between the United States and China in which the US largely exhausted its munitions inventories in as little as three to four weeks. Some munitions, anti-ship missiles among them, only lasted a few days. Replacing them would take years.¹

I remember stopping when I first read it. Out of munitions in less than a month? How could that be? I went back and read it again. How could the world's biggest military ever be on the verge of running out of munitions?

My first assumption was that this had to be about factories and outsourcing, and maybe some critical minerals. So, build more capacity, modernize the old plants, make more shells. Right?

That's partly right. It just isn't the real story.

To try and understand this better, in 2025 my team and I mapped the value chain that produces US and allied munitions, from raw materials through to battlefield use. We wanted to know how a munitions shortage could be possible, and what might be done about it. Then we analyzed it again last week, to see what may have changed since a year ago.

The First Hypothesis: Build More Factories

Our initial analysis (in 2025) confirmed what the Commission assessment had found months earlier. Manufacturing capacity really was constrained. Production of artillery shells, explosives and other munitions had been allowed to shrink for years, and the industrial base was struggling to expand at the pace suddenly being demanded of it.²

Had we stopped there, the conclusion would have seemed intuitive. Build more factories. Modernize existing plants. Increase production. However, value chain analysis rarely lets you stop where you'd like to. Every time we identified a constrained manufacturing step, we asked the same question.

What does this capability depend on?

At first the bottlenecks looked obvious. There weren't enough factories producing artillery shells.

So we asked what those factories depended on.

Propellants.

Then we asked what the propellants depended on.

Military grade nitrocellulose. Nitroglycerin. Specialized energetic chemistry.

So, we kept going, opening one nesting doll after another. Explosives depended on precursor chemistry produced by a handful of qualified suppliers. Primer production depended on chemicals and refining capabilities that existed in surprisingly concentrated markets. Some dependencies involved critical minerals. In several cases, the United States and its allies relied on supply chains in which China played a significant role, or on markets where a disruption at one or two suppliers could affect large portions of downstream production.³

Each answer pulled us one layer farther upstream. Again and again, what looked like the bottleneck wasn't the only bottleneck. It was the first of many.

Building more factories was still necessary. It just wasn't enough. Even the most modern production line can't manufacture propellant without military grade nitrocellulose, or produce insensitive explosives without qualified precursor chemistry. If those upstream capabilities remain fragile, downstream investment eventually runs into a wall.

At the time, I thought we'd reached the heart of the problem.

We still hadn't. There was a lot further to go.

Looking Again

Last week we ran it again. Since our first analysis of this value chain in 2025 the US and its allies have committed serious money to munitions production. It's ongoing; new investments continue to be announced. Existing facilities are expanding. Governments are paying much closer attention to supply chain resilience.

A lot of it is working. Downstream, where the money went first, load, assemble and pack capacity is no longer the tightest point in the artillery chain.⁴ Across the ninety capabilities in the current map, forty-seven are getting better.⁵

While nontrivial, many of those improvements turned out to be the low hanging fruit. What’s left are the problems that can’t be solved quickly, no matter how much money is spent. At least not right away.

Forty-three capabilities are still bottlenecked, and twenty-one of those severely. None of the twenty-one has an estimated duration under a year. Twenty carry timelines of three years or longer. Seventeen run five years or longer.

Fifteen of the twenty-one toughest bottlenecks sit in the first half of the chain: raw materials and energetic inputs, qualification, energetic synthesis and component manufacturing. Not one sits at final assembly.

I also started seeing the same names over and over.

Holston.

Radford.

Facilities such as Holston Army Ammunition Plant and Radford Army Ammunition Plant kept appearing, not because they assemble finished weapons, but because they produce capabilities on which much of the rest of the industrial base depends. Radford is the primary manufacturer and supplier of nitrocellulose and the primary supplier of solventless propellants.⁶ Holston makes RDX, HMX and the insensitive munitions explosives that go into artillery and bomb fills.⁷ Both are government-owned plants built in the 1940s and run under contract.⁸

These weren't ordinary industrial inputs that could simply be purchased elsewhere. Many required years to expand, years to qualify and years to integrate into production. Some depended on expertise accumulated over decades.

So the problem wasn’t factories. It wasn’t even mainly the supply chain. The industrial base was gradually rebuilding both.⁹ What remained looked, at first, like two separate problems.

Some inputs, such as military grade nitrocellulose, had few practical substitutes.

Others were trapped behind qualification.

It took us a while to understand that these weren’t really separate problems. They were often the same problem viewed from different points in the chain. Finding another material or supplier didn’t necessarily create usable supply. The substitute still had to be proven, accepted and qualified for the military system that would depend on it.

That turned out to be the real issue.

The Critical Distinction: Capacity vs Capability

Nearly all of the constraints that remained shared one characteristic. They required “qualification.”

I’ll admit I initially thought about qualification as paperwork. Necessary paperwork perhaps, but paperwork all the same.

I was wrong.

Imagine a scenario where several new companies begin producing military grade nitrocellulose. Intuitively, you’d expect the supply shortage to ease pretty quickly.

Except it doesn’t work that way.

Militaries don’t qualify broad categories like nitrocellulose, propellant or explosive. They qualify the specific material produced by a specific supplier, using a specific manufacturing process, for a specific military application. Another supplier’s version isn’t automatically a substitute, even if it’s chemically similar.

Before that substitute can become part of a military system, it has to repeatedly demonstrate that it performs exactly as expected. Not approximately. Not most of the time. It must ignite when it should, burn at the intended rate, remain stable during storage, behave consistently across production lots and continue performing after years in demanding environments.

An unreliable smartphone is inconvenient. An unreliable artillery shell, missile motor or detonator can cost lives. Qualification isn’t bureaucracy added after the engineering is done. It’s the process through which a new material, supplier, production method or facility becomes trusted military capability. That’s a high bar, and rightly so.

Finding a substitute is only half the problem. That substitute still has to be qualified.

Once I saw it this way, I understood why qualification appeared at every turn. Seventeen separate capabilities in our map were waiting for a product, process or facility to complete qualification, or for a national authority to accept the results.

That’s why building new munitions production capabilities in the US and allied nations is going to be slow. Reliability testing, certification and qualification take time. Investment can help, but it doesn’t always eliminate the time required to prove that a new material, process or supplier is ready for military use. Like Warren Buffett said, "No matter how great the talent or efforts, some things just take time. You can't produce a baby in one month by getting nine women pregnant.“¹⁰

Money can build another production line. It can’t instantly create years of successful testing, operating experience and confidence that a new material will behave exactly like the one it replaces.

There was another thing I had wrong, and I didn’t notice I had it wrong. I’d been picturing munitions as settled technology. A bullet is a bullet. I knew the missiles and the guidance and sensing were sophisticated, but say the word "munitions" and I pictured things that blow up. Without realizing it, I'd been imagining black and white WWII footage of wartime factories. Rosie the Riveter.

The field hasn’t stood still. Energetic chemistry keeps advancing. New insensitive explosives, new propellants, new manufacturing methods and new performance requirements continue to emerge. The goal isn’t simply to reproduce yesterday’s ammunition. It’s to build the next generation of ammunition that’s safer, more reliable and more effective. Every meaningful improvement has to be tested, validated and qualified before it becomes military capability. Qualification isn’t a one-time hurdle on a finished technology. It’s part of how an industry continues to innovate without sacrificing reliability.

That changed how I thought about munitions factories.

A factory is infrastructure. Capability is something different. Capability includes specialized chemistry, process knowledge, experienced workers, validated manufacturing methods, trusted suppliers, innovation and the accumulated operating experience required to produce military grade materials consistently enough to satisfy qualification.

The closest analogy I could think of was semiconductors. You could build an exact replica of a TSMC fab on American soil, but the building is only the beginning. Success depends on whether an organization can reproduce thousands of tightly controlled manufacturing steps with extraordinary consistency, supported by experienced people, qualified suppliers, a robust partner ecosystem and years of accumulated process knowledge.

The defense industrial base follows a similar pattern. The visible infrastructure matters. The (often invisible) capabilities operating within it matters more.

The Better Question

When I started this work, I thought I was investigating ammunition. Looking back, ammunition was the first clue. Every answer led us further upstream until the chain ended with qualified materials, validated processes, experienced people and institutions that had spent decades learning to produce these inputs reliably.

I also realized the goal wasn’t to recreate yesterday’s industrial base. It was to build the capabilities for tomorrow’s battlefield.

Somewhere along the way, the question changed. It wasn’t, “How do we build more factories?” It became, “How do we build new industrial capabilities?”

That progression changed how I thought about time. Building a factory is measured in years. Building an industrial capability is often measured in decades. Some of that knowledge can be documented. Much of it lives in experienced operators, qualified production methods and institutional memory.

Once those capabilities disappear, rebuilding them is far more difficult than replacing the buildings that once housed them. Creating new ones is harder still.

That’s why the progress of the past year matters more than the production numbers alone suggest. The easy constraints are clearing. The United States and its allies are restoring capabilities that had atrophied over decades while developing the next generation of energetic materials, manufacturing processes and weapons. The industrial base is recovering, but it isn’t returning to the past. It’s both rebuilding capabilities that time erased and creating the new capabilities the future will demand.

What it can’t do is hurry the part that’s left.

Could the United States really run out of ammunition in a month?

Perhaps, but now I think I see the challenge a little more clearly.

Capabilities like these take a long time to build and a short time to lose.

Endnotes

1. Commission on the National Defense Strategy, final report, July 2024, Chapter 7, "The Defense Industrial Base and Defense Production." The Commission was created by Congress in the FY2022 National Defense Authorization Act and released its report on July 29, 2024. The report attributes the three-to-four-week figure to unclassified public war games rather than to analysis the Commission conducted itself. https://www.rand.org/content/dam/rand/pubs/misc/MSA3057-4/RAND_MSA3057-4.pdf

   
   

2. Commission on the National Defense Strategy, final report, July 2024, Chapter 3, "Defense Industrial Base and Technological Innovation." The Commission attributes the shortfall to consolidation and underinvestment producing too few companies, workforce gaps, insufficient production infrastructure and fragile supply chains.

   
   

3. U.S. Geological Survey, "Minerals with Net Import Reliance on China" and "Antimony Statistics and Information," National Minerals Information Center. https://www.usgs.gov/centers/national-minerals-information-center/antimony-statistics-and-information

   
   

4. U.S. Army, "Army seeks to expand and accelerate 155 mm production," February 20, 2025. https://www.army.mil/article/283210/army_seeks_to_expand_and_accelerate_155_mm_production

   
   

5. Figures in this section, and the counts that follow, come from our own value chain analysis of US and allied munitions, first run in September 2025 and repeated in July 2026, together with a forensic comparison of the two.

   
   

6. Testimony of Brian Gathright, Vice President and General Manager, BAE Systems Ordnance Systems Inc., House Armed Services Committee, Subcommittee on Tactical Air and Land Forces, March 31, 2022. https://docs.house.gov/meetings/AS/AS25/20220331/114531/HHRG-117-AS25-Wstate-GathrightB-20220331.pdf

   
   

7. Same testimony. RDX is Research Department Explosive, HMX is High Melting Explosive, IMX is the insensitive munitions explosive family.

   
   

8. BAE Systems, "BAE Systems awarded U.S. Army contract to operate Holston Army Ammunition Plant," December 12, 2023. Both plants are government-owned, contractor-operated. BAE Systems Ordnance Systems Inc. has operated Holston since 1999 and Radford since 2012. https://www.prnewswire.com/news-releases/bae-systems-awarded-us-army-contract-to-operate-holston-army-ammunition-plant-302013330.html

   
   

9. On rebuilding upstream: U.S. Army, "U.S. Army awards contract for domestic TNT production," November 8, 2024; U.S. Army, "U.S. Army opens future domestic source of antimony sulfide," September 22, 2025; U.S. Department of Defense, "DOD Awards $192.5 Million to Establish Domestic Manufacturing Capabilities for Critical Defense Chemicals," January 11, 2024.

   
   

10. Warren Buffett, Berkshire Hathaway Inc. Chairman's Letter, 1985, in a passage on patience regarding the Capital Cities/ABC investment. https://www.berkshirehathaway.com/letters/1985.html