The factory capital stock
What did it cost to build all the world's factories? This second piece prices the ~10 million factories counted in the first — their total capital invested, net of depreciation — sets it against economic output, and traces the manufacturing capex supercycle: the four-decade wave of factory investment, and how completely China came to dominate its marginal dollar. The third piece then asks whether the AI buildout is a supercycle of the same order. Editorial order-of-magnitude estimates throughout; methodology in the appendix.
How much capital is standing in the world's factories?
A factory's value is its buildings plus its equipment, run down by wear and obsolescence. Add up every factory's structures and machines at replacement cost and the world's gross manufacturing capital stock is on the order of $34 trillion. Subtract accumulated depreciation — the wear already booked against those assets — and the net stock is roughly $24 trillion. That is the honest “what would it cost to rebuild what's standing, minus the wear” number.
Set against output, $24tn of net stock produces $16tn of value added a year — a capital-to-output ratio of about 1.5×. Each year the world spends roughly $3.7tn building and re-equipping factories (gross capex) while about $2.2tn of existing stock wears out — so net stock grows by the difference, a little over a trillion dollars a year.
Not all capital stock is equal
The same value added can sit on very different amounts of capital. The United States runs the leanest large factory base in the world — barely $2.8tn of net stock produces $2.9tn of value added, a ratio near 1×: old, heavily depreciated, but highly productive. China is the opposite — the newest and most capital-heavy base on Earth, roughly 1.84×, reflecting a decade of building capacity faster than it could be fully utilised.
| Economy | Net capital stock | Value added | Capital ÷ output | Note |
|---|---|---|---|---|
| China | $9.0tn | $4.9tn | 1.84× | Newest, most capital-heavy stock; utilisation runs below the West |
| World | $24.0tn | $16.0tn | 1.50× | |
| EU-27 & UK | $4.4tn | $3.0tn | 1.47× | |
| United States | $2.8tn | $2.9tn | 0.97× | Capital-light: older, heavily depreciated, high labour productivity |
The capex supercycle — China vs. the rest of the world
The reason the world's factory stock is so young, and so China-heavy, is a single generational wave of investment. For most of the twentieth century the rich world's manufacturing capex grew slowly. Then, from the late 1990s, China began investing in factories at a pace no economy had ever sustained — and kept doing it for two decades. The chart below splits world manufacturing capex into China and everyone else. The rest-of-world line barely doubles across thirty years; China's goes from a rounding error to roughly half of the world total.
Gross fixed investment in manufacturing, $ trillion per year, editorial order-of-magnitude estimates. The story is the shape: the rest-of-world line barely doubles across thirty years while China's goes from a rounding error to near-parity — the manufacturing capex supercycle in one picture.
This is the “supercycle”: not a boom-and-bust but a sustained, state-financed reallocation of a whole economy toward building productive capacity. It is why China now holds around 38% of the world's net manufacturing capital stock while producing about 30% of its value added — capital ran ahead of output. And it is the direct setup for the third piece: an investment wave this large is the natural yardstick against which to measure the AI capex boom.
What the numbers say
- The stock is huge but not bottomless. $24tn is about a fifth of one year of world GDP — factories are capital-heavy, but the world's total capital stock (housing, infrastructure, everything) dwarfs them.
- Depreciation is a slow tax. At ~$2.2tn a year, roughly a 9% annual burn, the world must invest over a trillion dollars a year just to stand still — before any growth.
- China over-built; the US under-built. The capital-to-output gap between them (1.84× vs 0.97×) is the whole geopolitics of manufacturing in one ratio.
Appendix — methodology
From investment to stock
Capital stock is built up by the perpetual-inventory method: accumulate past gross fixed investment, subtract depreciation at assumed asset lives (structures ~30–40 years, equipment ~10–15). Starting from world manufacturing gross fixed capital formation of ~$3.7tn/yr and a ~11-year effective average life gives a net stock in the low-to-mid $$20-trillions — the $24tn anchor used here. Gross stock (~$34tn) is before subtracting accumulated wear.
Capex series
The China-vs-rest-of-world series is an editorial reconstruction of manufacturing gross fixed investment, in nominal dollars, benchmarked to China NBS fixed-asset investment (manufacturing sub-total) and to World-Bank GFCF with a manufacturing share applied. The exact levels are soft; the shape — China's rise from ~5% to ~50% of the world total — is the robust finding.
Capital-to-output ratios
Ratios divide net manufacturing capital stock by manufacturing value added. The US figure (~1.0×) reflects OECD capital-stock data for a mature, heavily-depreciated base; China's (~1.8×) reflects a young stock built faster than utilisation caught up. These are the most defensible relative numbers on the page even if the absolute levels carry wide error bars.
Sources consulted
- UNIDO. World manufacturing value added, competitive-industrial-performance rankings.
- World Bank. Manufacturing value added (% of GDP), gross fixed capital formation.
- China NBS. Industrial enterprises 'above designated size', fixed-asset investment.
- OECD. Capital stock, consumption of fixed capital, investment by industry.
- IEA. Clean-tech and battery manufacturing investment.