US Defense R&D Spending Still Dwarfs Every OECD Rival

I keep coming back to one number that should make anyone who follows public budgets sit up a little straighter. For every $1,000 of economic output, the United States puts $3.64 into defense research and development. South Korea, the only other country that even looks serious on this yardstick, sits at $2.06. After that the floor drops. Britain is under a dollar. Germany is lower still. A cluster of wealthy European economies barely clear a quarter. Greece, at the bottom of a twenty-country ranking, spends nine cents.

What The Defense Research Gap Actually Looks Like

That is not a rounding error. It is a structural choice. Defense research is the part of the military budget that tries to buy the next decade rather than the current one. Missiles already in inventory do not count. Neither do salaries or fuel. This is money aimed at laboratories, prototypes, software architectures, materials science, and the ugly trial-and-error work that never makes a parade. When you scale it to the size of each economy, the American lead is almost theatrical.

Recent international comparisons covering 2024, or the latest available year, put the twenty-country average at $0.58 per $1,000 of GDP. The median is $0.27. Averages lie a little when one giant sits so far to the right of everyone else. Pull the United States out of the sample and the club looks modest, even timid. I have found that people argue about total military spending for years and still miss this slice. Total outlays tell you how large the force is today. Research intensity tells you how hard a government is trying to change what that force can do tomorrow.

The Ranking That Refuses To Flatten

Here is the picture in plain terms. The United States leads at $3.64. South Korea follows at $2.06. The United Kingdom comes third at $0.97. Germany is fourth at $0.79. France is fifth at $0.50. Then the list slides through Poland, Japan, Slovenia, Norway, the Netherlands, Australia, Sweden, Spain, Estonia, Türkiye, Finland, Latvia, Belgium, Hungary, and finally Greece at $0.09.

Look at the drop after second place. The American figure is nearly 1.8 times South Korea and more than 3.7 times Britain. It is more than seven times France and more than forty times Greece. That is not a gentle slope. That is a cliff.

Defense research intensity varies because countries do not treat the future of force the same way. Some buy platforms. Some pay people. Some bet on labs.

Absolute Dollars Still Matter More Than Ratios

Ratios are useful. They keep a small, rich country from looking lazy next to a giant. They also hide the raw scale. In absolute terms the United States remains the world’s largest military spender by a wide margin. One recent tally put 2025 outlays near $954 billion, roughly a third of global military spending that year. Even if the research share of that pile is only a slice, the slice is enormous compared with almost any ally.

There is a twist. Military spending as a share of American GDP eased in 2025 relative to the 2015–2022 average. That fact gets used in two opposite speeches. One camp says the burden is already high enough. The other says the share fell while threats did not. Both can quote a chart. Neither chart settles the research question. A country can trim operations and still pour money into the next generation of sensors, propulsion, and software. Or it can do the reverse and look busy while the laboratory pipeline thins out.

The Budget Request That Widens The Distance

The gap is not static. A fiscal 2027 defense request circulating in Washington called for $1.5 trillion, about 42 percent above the prior year’s funding level. Buried in that headline is a smaller but telling move: defense research would rise from $16.6 billion in 2026 to $18.7 billion in 2027, with officials talking up a general science and technology fund. Those are not the only research dollars in the system. They are the ones the request chose to highlight.

I’ve sat with enough budget tables to know that a request is not an appropriation. Congress will cut, shift, and relabel. Still, direction matters. When the largest spender asks for another step up in research while most peers remain under a dollar per thousand of GDP, the international ranking does not compress. It stretches.

Why South Korea Sits Alone In Second Place

South Korea is the only other OECD member above $2. That is not an accident of accounting. The country has spent years turning a threatened peninsula into an export machine for artillery, aircraft, and naval systems. Four of its firms showed up among the world’s top hundred arms producers by sales in 2024, and their combined revenues jumped about 30 percent from the year before. You do not get that kind of commercial burst without a long runway of public research and industrial policy.

Geography does some of the explaining. A neighbor with nuclear weapons concentrates the mind. Export markets do the rest. When domestic demand is real and foreign buyers are shopping, research stops being a prestige line item and starts looking like product development. Europe talks about strategic autonomy. South Korea has been practicing a harder version of it, with fewer speeches and more factories.

Europe’s Awkward Middle

Britain tripled net defense research spending between 2019 and 2024, from just over a billion pounds to a little more than three. That is a real acceleration. It still leaves the country under a dollar per thousand of GDP. Germany and France, the European Union’s two largest defense research spenders in absolute terms, sit at $0.79 and $0.50. Poland, which has been buying hardware at a sprint, is already near France on the research ratio. That detail is easy to miss if you only watch procurement headlines.

Perhaps the most interesting aspect is how quickly the European ranking falls after the first few names. Sweden and the Netherlands, countries with serious engineering cultures, land around $0.26 to $0.28. Spain is $0.23. Several Baltic and Central European states are lower still. These are not poor countries. They are countries that have chosen, for a mix of history, politics, and alliance cover, to let someone else fund the risky early work.

• Some governments prioritize ready forces over laboratories.
• Some treat research as a national industrial project.
• Some assume an ally will invent the next system and share enough of it.
• A few try to do all three and discover the budget cannot stretch.

There is no moral ranking in that list. There is only a trade-off. If you underfund research for a decade, you do not get a sudden miracle in year eleven. You get dependency, delayed programs, and a habit of buying what someone else already finished.

Civilian Spinoffs Are Real, And They Are Not Automatic

Defense laboratories have a long, messy record of leaking into ordinary life. Computing, aerospace, satellites, certain medical tools, and advanced materials all carry military fingerprints. The internet’s ancestors lived in funded networks. Jet travel borrowed from military propulsion. GPS began as a navigation aid for forces that could not afford to get lost. That history is not a slogan. It is a stack of accidents, contracts, and engineers who later changed jobs.

The link is not a vending machine. Secrecy slows diffusion. Classification walls keep papers unpublished. Export controls freeze components. A lab can spend a decade on a material that never leaves the compound. In my experience, people who treat every research dollar as a guaranteed civilian dividend are selling a story. People who treat every research dollar as pure waste are selling a different story. The truthful version is duller. Some projects migrate. Many do not. The ones that migrate often do so years later, after the original mission has moved on.

Military money can seed civilian breakthroughs. It can also lock knowledge behind a door and leave the rest of the economy waiting in the hallway.

That is why the American lead is not only a security fact. It is an industrial fact. A country that funds more early-stage work will, on average, produce more dual-use residue. Whether firms capture that residue depends on universities, venture capital, procurement rules, and the patience of program officers. The budget ratio is the starting gun, not the finish line.

How Countries Split The Same Pie Differently

Defense accounts are not one number. They are a set of competing appetites. Personnel costs rise with wages and benefits. Operations rise with deployments and maintenance backlogs. Procurement rises when fleets age. Research is the easiest line to cut in a tight year because the public cannot see a missing prototype. No parade features an experiment that failed on a Tuesday.

That political weakness explains part of the international spread. A government facing an immediate border crisis will buy ammunition before it funds a ten-year software stack. A government living under an alliance umbrella may decide the research can wait. A government that wants export orders will fund the labs that make those orders possible. Same pie. Different knives.

A rough way to think about national choices:
  Force today  — personnel, fuel, spare parts
  Force soon   — procurement of known systems
  Force later  — research, prototypes, science funds

The United States still spends heavily on the first two columns. What sets it apart on this particular ranking is the third. South Korea is the only peer that treats the third column as a core industrial strategy rather than a remainder. Everyone else, with a few partial exceptions, treats it as optional seasoning.

The Innovation Argument Has Limits

Fans of large research budgets like to recite the spinoff list. Critics like to recite cost overruns, canceled programs, and weapons that arrived late and heavy. Both lists are real. A high research ratio does not guarantee a smart force. It guarantees more attempts. Some attempts become standards. Some become museum pieces. The American system is famous for both outcomes in the same decade.

I’ve found that the more useful question is not whether research “works.” It is whether a country can afford to be wrong many times and still field something that matters. Iteration is expensive. Iteration under classification is more expensive. Allies who spend pennies on the ratio are not foolish. They are renting iteration from someone else, with all the political strings that rental implies.

Markets Notice Even When Voters Do Not

Investors who follow contractors already know the pattern. Research intensity feeds a pipeline of programs, and programs feed backlog. When a government signals a multi-year lift in science and technology accounts, prime contractors and specialist suppliers start hiring systems engineers before the first milestone review. That is not romance. That is working capital.

The same signal travels into dual-use corners: satellite buses, secure communications, advanced manufacturing, certain medical imaging tools, and materials that later show up in civilian aircraft. Not every ticker benefits. The ones closest to the laboratory door usually move first. The ones that only assemble mature hardware move later, if at all.

Is that a reason to cheer every extra dollar? No. Public research can crowd out other priorities. It can also misallocate talent toward classified work that never commercializes. A serious reader holds both thoughts at once. The ranking still stands. America is funding the attempt at a scale almost nobody else will match.

Allies, Burden Sharing, And An Old Argument

Every few years the burden-sharing debate returns. It usually focuses on the headline share of GDP spent on defense, not the research slice. That is a mistake. A partner can hit a spending target by raising pay, buying off-the-shelf vehicles, or restocking magazines. Those steps help readiness. They do not rebuild a design base. If the alliance wants more than mass, someone has to fund the unglamorous early work.

Right now that someone is mostly Washington, with Seoul as the only other rich-country government clearly in the same conversation. London has accelerated. Berlin and Paris spend meaningful absolute sums because their economies are large. The ratios still tell a quieter story. Most of the club is content to remain a customer of tomorrow’s architecture rather than an author of it.

1. Measure readiness with personnel and munitions.
2. Measure modernization with procurement of known designs.
3. Measure future options with research intensity.
4. Do not pretend those three meters are the same instrument.

What A Wider Gap Would Mean In Practice

If the 2027 request survives in anything like its proposed shape, the American lead grows again. Not because allies collapse, but because they move slowly while the baseline in Washington jumps. A 42 percent lift in the overall account, even if research only rises by a couple of billion, changes the conversation inside laboratories. Program managers stop asking whether a risky prototype can be funded and start asking which risky prototype goes first.

That sounds abstract until you translate it into timelines. Software-defined radios, uncrewed systems, missile defense sensors, hypersonic experiments, and space awareness tools all live or die on multi-year research continuity. Stop-start funding kills teams. Continuous funding keeps the tacit knowledge in the building. Allies who fund in fits and starts discover that their best engineers take jobs elsewhere. Then they buy the finished kit and wonder why it costs so much.

The Quiet Cost Of Being The Laboratory

There is a less comfortable side. Being the country that funds the early work means absorbing the failures. It means paying for dead ends that never appear in export brochures. It means carrying the political heat when a program slips. Allies who spend $0.20 per thousand of GDP get to criticize the overruns without funding the attempts that produced them. That arrangement can last a long time. It lasts until the laboratory country gets tired, or until the customer countries decide the product line no longer fits their politics.

I do not think fatigue is visible in the current request. If anything, the opposite is true. The request treats research as a lever, not a luxury. Whether that stance survives committee rooms is another matter. Budget seasons are long and unkind.

Reading The Median, Not Just The Champion

The median of $0.27 is the number I would tape to a cabinet wall. It says that a typical rich country in this sample treats defense research as a rounding item. The average of $0.58 only looks healthier because the United States and South Korea drag it upward. Policy debates that lean on the average will overstate how common this investment really is.

That matters for anyone who assumes “the West” funds science the same way. It does not. The West, on this measure, is a long tail with two spikes. If you design alliance strategy around the spikes, you will be surprised by the tail. If you design it around the tail, you will underuse the spikes. Neither error is theoretical. Both show up in capability gaps that take a decade to close.

A Few Practical Takeaways Without The Speech

First, ratios beat slogans. Asking whether a country “supports innovation” is theater. Asking what it spends per thousand of GDP on defense research is a measurement. Second, South Korea is the comparison that should make European finance ministries uncomfortable, not because Seoul is perfect, but because it converted research into exportable industry on a timetable others only describe. Third, civilian benefits exist and they are leaky. Plan for leakage. Do not budget as if leakage were a dividend check.

Fourth, a larger American request does not settle whether the money will be spent well. It only settles the scale of the attempt. Oversight still matters. Program discipline still matters. The ranking does not grade quality. It grades intensity.

Fifth, markets and ministries should stop treating research as a footnote to procurement. The footnote is where the next procurement list is born. Ignore it and you will keep buying last year’s architecture at next year’s prices.

Where This Leaves The Reader Who Does Not Live Inside Budgets

You do not need a spreadsheet addiction to use this ranking. If you care about technology, notice where the early money sits. If you care about alliances, notice who is writing the next manual and who is waiting for the translation. If you care about public finance, notice that research is the line item with the longest lag between payment and proof. That lag is why it gets cut. It is also why cutting it is so expensive later.

Will the American lead shrink? Not on current published intentions. Could allies close part of the gap with a few determined budget cycles? In principle, yes. Britain already showed that a tripling is possible when politics align. Most others have not tried that experiment. Until they do, the map stays lopsided: one very large laboratory, one serious runner-up, and a crowded field of careful spenders who hope the future arrives ready-made.

That hope is a strategy. It is just not the same strategy as $3.64 per thousand of GDP. The difference will keep showing up in patents, prototypes, export catalogs, and the quiet argument over who actually pays to invent the next system. I would rather watch that argument in the open than pretend the ratios are close enough. They are not. Not even a little.