Rearmament: real, and starting from the lowest base since the Cold War

Status: drafted 2026-08-13 Reachability swept first. OWID's military-expenditure-share-gdp 404s; its live slug is military-spending-as-a-share-of-gdp-sipri (recovered by following the 301). That slug carries 167 countries and no World or regional aggregate — the one line this argument turns on cannot be drawn from it. The World Bank republishes the same SIPRI series with a World aggregate (MS.MIL.XPND.GD.ZS), one year behind. Aggregate beats extra year; the World Bank connector is the source. Vintage checked: World Bank lastupdated: 2026-07-13, data through 2024. SIPRI's own 2025 release exists (OWID has it) — see open question 1.

Segment: The peace dividend was enormous, it ended in 2018, and the rebound is smaller than it sounds

• Source theme(s): No exact bank theme — adjacent to Batch 24 #1 (the warhead decline), and this is its economic twin. Recorded honestly as a gap the same way the shipped-artifact back-map does.
• Spine role: main story
• The framing people carry: The world is rearming at frightening speed; Europe is scrambling to rebuild armies it let rot; a new arms race is on.
• Hypothesis under test: That "record military spending" means the world is returning to Cold War levels of militarisation.
• Verdict: QUALIFIED — the direction is right and the magnitude is badly wrong. Rearmament since 2018 is real and in places dramatic. But the world spends 2.47% of GDP on its militaries against 6.14% at the 1962 peak and 3.32% in 1990. Every headline about record spending is measured in dollars against a world economy several times larger. As a share of what we produce, 2024 is 59% below 1960 and still below every year of the Cold War.

The numbers (every row sourced — no exceptions)

Source: SIPRI via World Bank Open Data, indicator MS.MIL.XPND.GD.ZS, accessed 2026-08-13. Tier A (existing World Bank connector).

The texture (what narrative flattens)

• The lead finding: the denominator. "Record military spending" is a dollar claim about a much larger world economy. On the share-of-GDP measure the world is at 2.47%, against 6.14% in 1962 and 3.32% at the end of the Cold War. The rebound is real — up 16.7% from the 2018 floor — and it starts from the lowest militarisation the modern record contains.
• The turn is dateable: 2018. Not 2022. The world trough is 2018 and the line was already rising before the full-scale invasion of Ukraine. That ordering is the same shape as the nuclear story — the direction changed before the event everyone attributes it to.
• Two of the biggest movers are not who you'd guess. Poland now spends a larger share of its economy on defence than the United States — 4.15% to 3.42% — having been at 1.75% in 2008. Japan has broken the 1%-of-GDP norm it held for half a century, reaching 1.37%.
• Germany is the story people think they know, and the numbers are smaller than the rhetoric. From 1.12% in 2015 to 1.89% in 2024: a 69% increase, and still below the 2.51% it spent in 1990 and well below 1963's 4.88%.
• China is the counter-example that matters. The arsenal growing fastest in warheads (Batch 24 #1: 350 → 620 since 2020) is spending a falling share of GDP — 2.45% in 1990, 1.61% in 2021, 1.71% now. Its absolute spending has grown enormously; its economy grew faster. Both facts are true and the share measure is the one nobody quotes.
• Russia is the genuine outlier, at 7.05% and nearly double its 2021 share — a war economy, and visible as one in a single line.

Datoid artifact

• Dashboard military-spending, 4 panels: the world arc 1960→2024; the two Cold War poles; the European rearmers (Germany, Poland); and China vs the world. security category, tag security.
• Pairs directly with the nuclear-arsenals dashboard — same domain, opposite measure, and the China divergence is the bridge between them.

Story justification (Stage 1 rule 5)

• Reveal that needs sequencing? NO → skip.
• Because: the argument is a single-line correction — put the share-of-GDP series on screen and the "return to Cold War levels" claim dies immediately. There is no withheld second dataset; the composition detail (Poland, China) is a set of parallel comparisons a panel shows in one view. A story would restate the dashboard.

Open questions / where I'm guessing

These must NOT become spoken assertions.

1. SIPRI's 2025 figures exist and this series stops at 2024. OWID carries SIPRI to 2025 per-country. Do not quote any 2025 number from this dashboard, and do not say "the latest year" without saying 2024.
2. Share of GDP is one denominator among several and it is not a measure of military capability. A country can hold spending flat as a share while its economy — and so its absolute budget — doubles. Say "share of the economy" explicitly every time; the whole point of the segment collapses if a listener hears it as "spending fell".
3. China's official defence budget is widely held to understate actual spending; SIPRI publishes an estimate that adjusts for this, and I have not verified how much. The direction (share falling while the economy grows) is robust; a precise Chinese percentage is not. Do not lean on 1.71% as exact.
4. No causal claim is supported here. Whether rearmament is a response to Russia, to US posture, or to domestic politics is not in this data.
5. 1960 is the series start, not the Cold War peak of militarisation. Some countries were far higher in the 1950s. "Since 1960" is the honest phrasing.