Industrial Revolution Atlas

2026 Documented phases active

Open a country profile

Coal-led Mixed Coal-light / non-coal-led Path not established

Map boundary geometry: Natural Earth via world-atlas 2.0.2.

Small-state inset

Every entry without a reliably tappable boundary remains selectable here.

What is missing?

Industrialization needs more than electricity access. It needs enough power per person for production, enough manufacturing output to sustain supplier networks, and a grid reliable enough that factories do not finance a second power system behind the meter.

This view compares 59 countries whose evidence does not establish a broad manufacturing take-off. Values use the latest available 2022 to 2024 observations. They are normalized per person so countries of very different sizes can be compared on the same scale.

AccessWho can connect. It does not measure whether power stays on.
kWh / personHow much electricity the economy can actually use.
MVA / personManufacturing value added, a direct measure of industrial depth.
Firm powerContinuous, stable supply needed by factories and data centres.

LOW-ENERGY BASE

Before take-off

<500 kWh / person

Electricity is usually concentrated in household lighting, with little headroom for production.

INDUSTRIAL TAKE-OFF

Factories emerge

500 to 2,000 kWh

Manufacturing begins to absorb power at scale. MVA commonly moves through roughly $200 to $600 per person.

DEVELOPED INDUSTRIAL BASE

Power is abundant

4,000+ kWh / person

No country below 4,000 kWh per person reached HDI 0.9 in Pasternak's cross-country study. Treat this as a floor, not a verdict. Source

AI INFRASTRUCTURE

A new kind of load

100+ MW, continuously

Hyperscale AI facilities start near 100 MW. Tier III design targets 99.982% availability, about 1.6 hours of downtime a year. IEA · Uptime target

Inspect a country

The bars share fixed ceilings: 100% access, 4,000 kWh per person, and $2,000 manufacturing value added per person.

All 59 country gaps

Sorted by electricity use per person

Filter metric

Minimum

Maximum

Compare countries side by side

Same definitions, same scales, no rank score.

Country A

versus

Country B

Aligned timelines

Both countries use the same 1750 to 2030 scale. Select a phase bar to open its evidence and sources. The overlap row marks years when both records show a documented industrial phase. Context bars are excluded.

Country timelines

Shared scale: 1750 to 2030. Select any bar for detail.

Country index

Eight charts that explain the industrial divide.

The evidence, then the argument. Each chart is built from the atlas research corpus and makes one claim. The final section looks forward, with every forecast explicitly labeled as inference rather than fact.

Bar chart showing the number of countries beginning a first industrial phase in each decade, with the 1900s as the peak. — Chart 01 · Industrial diffusion
More than half of all national industrial take-offs began after 1900, and the single busiest decade was the 1900s, when 43 countries started. Industrialization is not an 18th-century story with a long tail; for most of the world it is a 20th-century event, recent enough to sit inside living memory. The three pre-1750 records are colonial-era entries, flagged separately, not industrial phases.
Source: Industrial Revolution Atlas country-phase research (Sept 2026).

Scatter plot comparing electricity access with electricity use per person across 59 countries. — Chart 02 · Access versus use
Access counts the connection; kWh per person counts what the connection can run. Afghanistan reaches 87.8% of its population yet uses only 160 kWh per person, while Palau's 100% access delivers 6,713 kWh per person. At the bottom, Chad's 13.4% access and 19 kWh per person show what the access metric hides: a wire with almost nothing behind it.
Source: World Bank WDI / Our World in Data via the Atlas 59-country gap study (Sept 2026).

Log-scale scatter plot comparing electricity use with manufacturing value added per person across 52 countries. — Chart 03 · Power and manufacturing
Across 52 countries, the quadrant that would mark an industrialized economy, abundant power plus a real manufacturing base, is empty. The closest case is Equatorial Guinea at $1,808 of manufacturing value added per person, and even it sits far below both reference lines. Palau shows the mirror failure: 6,713 kWh per person with only $219.90 of manufacturing value added, power without industry.
Source: Atlas 59-country gap study; 7 countries lack MVA data (Sept 2026).

Ranked bar chart of the 15 largest electricity-use shortfalls against a 4,000 kWh per-person reference. — Chart 04 · Electricity shortfall
The fifteen largest gaps are all above 3,870 kWh per person short of the abundant-power reference, which means the worst-off countries use roughly 1 to 3 percent of what an industrialized economy consumes. Chad, the largest gap, uses 19 kWh per person in a year: about half a percent of the reference. These are not marginal deficits; they are a different energy universe.
Source: Atlas 59-country gap study; shortfall = max(0, 4,000 minus kWh/person) (Sept 2026).

Bar chart counting countries by their dominant source of electricity generation. — Chart 05 · Dominant electricity source
114 of 186 countries still get most of their electricity from coal, oil, or gas, and hydro remains the single most common number-one source at 53 countries. Only nine countries worldwide are led by wind or solar. Any plan that assumes the developing world's industrialization will be clean from the start is arguing against the current scoreboard.
Source: Ember / Energy Institute via Our World in Data, via the Atlas energy-mix data (Sept 2026).

Bar chart grouping 191 countries into four electricity-use tiers per person. — Chart 06 · Electricity admission tiers
89 of 191 countries sit below 2,000 kWh per person, the level where factories reliably emerge, and 46 sit below 500. Only 63 countries clear the 4,000 kWh abundant-power tier. The AI buildout is landing on a planet where most countries cannot yet power a factory, let alone a data center.
Source: World Bank WDI / Our World in Data via Atlas indicators (Sept 2026).

Bar chart grouping 59 countries by their primary binding constraint to industrialization. — Chart 07 · Binding constraints
There is no single bottleneck: the largest category, grid and transmission fragility at 14 countries, covers less than a quarter of the set. Generation, import dependence, conflict, and broken utility finances each bind a meaningful share. Any strategy that treats the power gap as one problem with one fix will miss most of it.
Source: Atlas 59-country gap study, categories derived from gap notes (Sept 2026).

The 4,000 crossing rule

Institutional milestones behind grid depth. High-consumption paths and access or integration precursors are labeled separately so a connection program is never mistaken for crossing the consumption threshold.

Britain1926–1938Statutory Central Electricity Board, common carrier, central dispatch High-consumption path

United States1935–1960REA long-tenor federal loans to rural cooperatives High-consumption path

Ghana1961–1965Akosombo construction and VALCO anchor-load agreement Anchor-load milestone

South Korea1965–1987Statutory rural program, state-directed credit, KEPCO tariff base High-consumption path

Vietnam1995–2013EVN national utility, tariff pooling, sequenced grid extension Access precursor

India1991–2013PGCIL assets, Grid-India dispatch, national synchronous grid Integration precursor

Brazil2003 onwardSector levies and payment after verified connections Access precursor, reported, verify

China2000–c.2015State grid monopolies, common standards, national cost pooling Approx. endpoint

France–Spain2008–2015INELFE asset company, equal TSO equity, EU grant, EIB debt Integration model

Kenya2015 onwardDensification and fee finance through the incumbent utility Access precursor, reported

Industrial Revolution Atlas by Richardson Dackam

Chart 08 · Institutional paths

The evidence does not support a shortcut from connections to abundant reliable power. The documented path combines long-duration capital, a statutory planner or operator, a party carrying risk, and a revenue institution. Exact threshold-crossing years are not consistently available, so the chart shows cited institution-building windows and marks approximate endpoints rather than inventing precision.

Sources are linked point by point inside the chart. Inference from the grid-building evidence review, Sept 30, 2026.

What the evidence suggests next

Inference, not fact. The numbers above are measured; everything below is a projection built on them, and it should be read as a set of argued hypotheses, not as predictions with confidence intervals.

The next decade

The diffusion chart carries the essay's starting fact: 107 of 196 industrial take-offs began after 1900. Industrialization is fast when the conditions are right, which means the next decade's question is not whether latecomers can industrialize but what is binding them right now. The constraint chart says the answer is fragmented: grids, generation, fuel imports, conflict, and utility balance sheets, in roughly that order. Inference: the countries that move fastest in the 2026-2036 window will be the ones where the binding constraint is narrow and fixable, which points to the import-dependence group (10 countries) and the finance group (6 countries). A country that imports its fuel or power has a procurement problem, and a country with broken tariffs has a policy problem; both are solvable on a five-year horizon in a way that civil conflict is not. The conflict-bound seven are, on this reading, unlikely to close their power gaps within the decade regardless of capital availability, because the constraint is not capital. The generation-shortfall group (11 countries) is the swing category: firm power takes years to build, so the decade's winners and losers will largely be decided by which of these countries started building in the early 2020s. Watch announced capacity, not announced targets.

The future of AI

Here is the mechanism the charts imply. Hyperscale AI needs a continuous 100 MW block of firm, cheap power, and 89 of 191 countries cannot yet deliver 2,000 kWh per person, let alone a dedicated industrial load. Inference: the AI century will deepen the power divide before it closes it, because data centers will cluster where power is already abundant and grids are already strong, which is the same 63-country tier that already industrialized. That concentration has a second-order effect the mix chart makes visible: 114 of 186 countries are fossil-led, so the marginal megawatt for AI in most of the world is a fossil megawatt. The "clean AI" narrative holds only inside the countries that already have clean grids; everywhere else, inference suggests AI demand will first extend the life of fossil generation, because it is the only firm capacity available on the relevant timescale. The optimistic counter-case, and it is genuinely a hypothesis rather than a forecast: solar-plus-storage is now cheap enough that the access-gap and too-small-to-industrialize countries (8 of the 59) could skip the grid-centric path the way they skipped landlines. If that happens, it will show up first in the mix data as rising solar-led counts, and the chart to watch is chart 5's right tail.

Nations at risk

Combine the shortfall ranking with the constraint categories and three risk profiles emerge, all inference. First, the conflict-bound states with the largest shortfalls, South Sudan, Yemen, Somalia, the Democratic Republic of the Congo: their power gaps are symptoms of state failure, and no energy investment thesis should treat them as energy problems. Capital deployed there is exposed to the conflict constraint, not the power constraint. Second, the import-dependent states: their risk is price exposure, and the next oil or LNG shock reprices their entire industrialization path overnight. Third, and most interesting, the grid-fragility group (14 countries, the largest category): these countries may generate enough power on paper while losing it in transmission, which means headline capacity numbers overstate their real position and investors who screen on installed megawatts will misprice them. The meta-risk, the one the atlas as a whole points at: the world is about to run two industrial revolutions at once, the AI buildout in the power-rich countries and the basic-electrification catch-up everywhere else, and they will compete for the same turbines, transformers, engineers, and capital. Inference: equipment bottlenecks, not policy, may be the binding constraint of the late 2020s, and the countries with the weakest procurement muscle, which the constraint chart identifies, will lose that bidding war first.

What is measured, and what remains uncertain

The index includes all 195 sovereign states, plus Taiwan. All 196 entries now have a researched timeline and energy-path assessment. Every entry carries two to four country-level references and an explicit evidence-confidence grade. The 139-country expansion and the original 57-country foundation were checked in separate citation passes so readers can audit the underlying chronology and energy claim.

UNIDO's 2026 classification covers 212 economies and territories: 64 industrial and 148 industrializing. Its public announcement does not expose the full annex in a reliably extractable form, so only Costa Rica receives a verified status here.

The five gold watchlist markers identify Bangladesh, China, India, Indonesia, and Viet Nam, the Emerging Industrial Economies named in UNIDO's public announcement. That label means manufacturing momentum, not “almost industrialized.”

The Gap view

The Gap view covers 59 countries where the evidence does not establish a broad manufacturing take-off. It normalizes electricity access, annual electricity use, and manufacturing value added per person. Figures mix the latest available 2022 to 2024 observations and are indicative, not synchronized national accounts.

The 4,000 kWh and $2,000 MVA lines are comparative reference bands, not pass/fail definitions. Access does not measure reliability. The AI threshold is different again: it asks whether one site can receive at least 100 MW continuously, not how national electricity divides across the population.

The energy-path rule

Coal-led: coal was at or near the largest sustained primary-energy source during take-off.
Mixed: coal mattered but did not dominate, or imported coal shared the work with hydro, oil, gas, or nuclear.
Coal-light or non-coal-led: little or no domestic coal, so industrialization relied on hydro, imported fuels, oil and gas, nuclear, geothermal, wind, solar, or institutional adaptations. This does not mean fossil-light: Saudi Arabia and Iran are oil-and-gas-led cases.
Path not established: the evidence does not establish a manufacturing take-off, so absence of coal cannot be treated as proof of coal-light industrialization.

Coal-light does not mean zero coal, and it does not mean fossil-free. Sweden imported coal before hydro displaced it, while Uruguay imported roughly 92 percent of its coal from Britain during 1890 to 1911. The label captures the missing domestic coal engine.

Phase dates

Each phase is an interpretive, rounded interval. The bars are built to reveal sequence and compression, not to declare a single official “industrial revolution year.” A dagger marks the five phase interpretations where the literature or political record is explicitly contested. Hatched bars marked Context record colonial or other external-rule conditions. They are not counted as the territory’s sovereign industrial phases.

Evidence confidence

High means multiple corroborating authoritative sources and a well-documented trajectory. Medium means real sources support the sequence, but some dates or energy shares rest on secondary summaries. Low marks thin records, especially microstates and post-conflict states. Across the atlas: 61 high, 109 medium, and 26 low. Confidence measures the evidence base, not a country's industrial performance.

Research sources

Expanded and validated September 23, 2026