Human Exposure to the Global Polycrisis, 2025–2100

In 2025, 2.0 billion people live in the 44 countries that INFORM classes as high or very high risk: 0.83 billion in the 16 very-high-risk countries, from South Sudan and Somalia to Afghanistan, Yemen and Haiti, and 1.2 billion in the 28 high-risk countries, which include Pakistan, Bangladesh, Myanmar, Kenya and Uganda. The 33 countries that sit in the top third of all three INFORM dimensions at once — hazard, vulnerability and lack of coping capacity — hold 1.5 billion people, 18.5% of the world.

What sets these countries apart demographically is youth. The median age in high- and very-high-risk countries is 19, half that of low- and very-low-risk countries at 38. Because their populations are young, they account for a much larger share of births and children than of people: 42.8% of the world's 132 million births in 2025, and 36.6% of its 2 billion children under 15, against 24.6% of its population. Exposure today already falls disproportionately on the young.

The map lets the layers be compared. The INFORM risk layer shows the familiar belt from the Sahel through Central and East Africa to the Horn, the Middle East and South Asia. The population-change layer shows where growth to 2100 will be fastest, and it overlaps that belt almost exactly: the Democratic Republic of the Congo, Niger, Chad, Somalia, Angola and Tanzania are each projected to grow between 3.2 and 3.8 times by 2100. The median-age layer shows the same geography from the other side.

The simplest projection holds every country's 2026 risk class constant and lets population change as the UN projects. It is not a forecast — risk will change — but it isolates the effect of demography. On this basis the population of today's high- and very-high-risk countries grows from 2.0 billion in 2025 to 3.0 billion in 2050, 3.9 billion in 2075 and 4.4 billion in 2100. Their share of humanity rises from 24.6% to 31.6% in 2050 and 43.3% in 2100, while the world population first grows and then, after the 2080s, begins to fall.

The most severe classes grow fastest. The population of the 16 very-high-risk countries rises from 0.83 billion to 1.39 billion in 2050 and 2.21 billion in 2100, from 10.2% of humanity to 21.8%. The compound core — the 33 countries in the top third on every INFORM dimension — grows from 1.52 billion to 3.72 billion, from 18.5% to 36.6% of the world. Meanwhile the population of today's low- and very-low-risk countries, which are mostly older and richer, is already at its peak: it shrinks from 2.85 billion to 2.12 billion by 2100.

This is demographic momentum. Populations with many young people keep growing for decades even after fertility falls, because the large cohorts already born are entering childbearing age. It cannot be reversed quickly by any policy, which is why it should be treated as a given in planning. The question it poses is not whether these countries will grow but whether the risk they face will fall fast enough for growth to be a dividend rather than an exposure.

The regional pattern sharpens over time. In 2025, sub-Saharan Africa holds 1.07 billion of the 2.0 billion people living in high- and very-high-risk countries, 53%. South Asia holds 0.50 billion (25%), mostly in Pakistan, Bangladesh and Afghanistan, and the Middle East and North Africa 0.23 billion (11%). Latin America, East Asia and Europe together hold under a tenth. By 2100, with today's risk map, sub-Saharan Africa's exposed population almost triples to 2.95 billion and its share rises to 67%, while South Asia's grows to 0.88 billion but its share falls to 20%.

Individual countries drive much of this. Among today's high- and very-high-risk countries, the largest populations in 2100 are projected to be Pakistan (511 million), Nigeria (477 million), the Democratic Republic of the Congo (431 million), Ethiopia (367 million) and Tanzania (263 million), followed by Bangladesh, Angola, Sudan, Afghanistan, Uganda, Yemen and Mozambique. The Democratic Republic of the Congo alone is projected to grow nearly fourfold, from 113 million to 431 million.

Regional concentration has a practical meaning for anyone planning resilience investment. Today's international architecture for crisis response, adaptation finance and development assistance was built when exposure was more evenly spread. By the second half of the century, the people it needs to reach will be overwhelmingly in Africa and in a handful of very populous countries in South Asia. That is where the gap between need and capacity, measured in section 1, will matter most.

Following births rather than people shows exposure moving down the generations. Attributing every birth to its country's 2026 risk class, the share of the world's births occurring in today's high- and very-high-risk countries was 19% in 1960 and 26% in 1990. It is 42.8% in 2025, and the UN projections take it to 48% in 2050 and 53% in 2100. Over 2025–2100 as a whole, 4.7 billion of the 9.5 billion projected births — 49% — will be in countries that are high-risk today. The share in very-high-risk countries alone rises from 21% of births today to 28% by 2100.

Climate science reaches the same conclusion from a different direction. Grant, Thiery and colleagues, writing in Nature in 2025, estimated that 52% of children born in 2020 will face lifetime heatwave exposure that would have been virtually impossible without climate change if warming is held to 1.5°C, rising to 92% at 3.5°C. Under current policies, around 2.7°C, the share of each cohort facing such exposure at least doubles between those born in 1960 and those born in 2020, and it is highest for the most socioeconomically vulnerable children.

The two findings compound. The children most likely to face unprecedented climate extremes are increasingly born in the countries least able to protect them, and they are being born in larger numbers there than anywhere else. UNICEF's 2024 State of the World's Children projects about eight times as many children exposed to extreme heatwaves in the 2050s as in the 2000s. The explorer below follows a single birth year through the century.

Heat is the hazard that most directly links climate change to where people live. Using the 1991–2020 climate, without any further warming, the share of humanity living in countries whose national mean temperature is 25°C or more rises from 18.8% in 2025 (1.54 billion people) to 25.2% in 2100 (2.56 billion), because the hottest countries are among the fastest-growing. The overlap with high risk grows too: the population of countries that are both hot and high-risk doubles, from 0.78 billion to 1.57 billion.

Warming adds to the demographic shift. Lenton and colleagues, writing in Nature Sustainability in 2023, defined a "human climate niche" from the temperatures at which human populations have historically been densest. They estimated that warming has already pushed about 9% of people, more than 600 million, outside it, and that current policies leading to around 2.7°C could leave between 22% and 39% of the end-of-century population outside it. Limiting warming to 1.5°C rather than 2.7°C would mean about five times fewer people exposed to such dangerous heat in 2100.

The young are especially sensitive. Infants and small children regulate body temperature less effectively than adults, and heat raises risks in pregnancy. Heat interrupts schooling, and it reduces the working hours and incomes of the parents who support children. Because the hottest, fastest-growing countries are also the youngest, heat exposure is concentrated in childhood to a degree that national averages conceal. The mean-temperature measure used here is a national, area-weighted average that understates heat where people live in cool highlands and overstates it in empty deserts. It is a guide to direction, not a count of heatwaves.

The frozen-risk baseline is only one possibility. Risk can rise or fall, and section 6 of this series' geography of global vulnerability report showed how much of it is within reach of policy. The scenario tool below recomputes every country's risk with INFORM's own formula under four paths and applies the UN population projection. The frozen path holds 2026 risk constant. The trend path extends each country's 2017–2026 change in risk, which has risen on average, and about four times as fast in high-risk countries (0.04 points a year) as elsewhere (0.01). The peer benchmark path moves each country's vulnerability and coping capacity to the best quartile of its income group by a chosen year. The peer plus peace path also halves conflict hazard.

The results span a wide range. If recent trends continued, the exposed share would reach 36% by 2050 and 53% by 2100, with 2.9 billion people in very-high-risk countries — a divergence that is speculative over 75 years but shows where current momentum points. If every country reached the peer benchmark by 2050, the exposed share would fall to 16% in 2050 and no country would remain very high risk; but population growth in the countries that stay high-risk would push the share back up to 24% by 2100. Only when conflict also halves does exposure stay low, at 7% in 2050 and 11% in 2100.

Two lessons follow. Speed matters: reaching the benchmark by 2075 instead of 2050 leaves exposure at 27.5% in mid-century, higher than today, because population grows faster than risk falls. And a one-off improvement is not enough, because the countries growing fastest keep growing after the benchmark is reached. Resilience has to keep pace with population for the whole century. Conflict is the multiplier that decides whether improvements in vulnerability and coping translate into lower exposure.

The analysis points to priorities that differ from those suggested by hazard maps alone. The first is to plan for the population that is coming. Health systems, schools, early warning, water and social protection in today's high-risk countries need to be built for populations that will be 50% larger by 2050 and more than twice as large by 2100, and much younger than anywhere else. Systems sized for today's population will be overwhelmed by tomorrow's even if the hazards stay the same.

The second is to reduce risk faster than population grows. The scenarios show that a one-time improvement is not enough; resilience has to compound. That favours investments that scale with population: institutions that can deliver services to growing cities, social protection systems that expand automatically in shocks, and early warning that reaches every new household. The third is peace. Conflict hazard is the factor that decides whether improvements in vulnerability and coping translate into lower exposure, and it is the factor that most changes the 2100 outcome.

The fourth is to centre children. The people most exposed this century are mostly not yet born, and those already born are young. Child-sensitive measures — heat protection in schools and health facilities, nutrition and immunisation that hold during crises, birth registration and identity that let social protection reach families — are therefore resilience measures, not only welfare measures. The fifth is mitigation: every tenth of a degree avoided reduces the number of people pushed outside the human climate niche, and most of those people will live in the countries highlighted here.

Risk is from the INFORM Risk Index Mid-2026 (v0.7.3, European Commission Joint Research Centre) for 191 countries, using INFORM's class thresholds. Population, births, children under 15 and median age for 1950–2100 are from the UN World Population Prospects 2024 (medium variant from 2024) as published by Our World in Data; the 191 countries cover 99.5% of world population. Exposure counts sum national populations by risk class, so everyone in a high-risk country counts as exposed; they do not capture risk within countries or migration beyond what the UN projections include.

The frozen path holds 2026 risk constant. The trend path extends each country's least-squares slope in INFORM risk over 2017–2026 (trend edition), capped at 0 and 10. The peer paths move each vulnerability and coping component to the 25th percentile of the country's World Bank income group, recompute risk with INFORM's aggregation formula (verified to within 0.05 of published scores), apply the change to the published 2026 score, and interpolate linearly from 2026 to the chosen year; the peace variant also halves human hazard. Cohort shares attribute births to the country's 2026 class.

Heat uses the 1991–2020 area-weighted mean of ERA5 monthly surface temperature (Copernicus, via Our World in Data); no warming is added to the population counts, and literature figures describe warming. All projections are scenarios, not forecasts. The map uses Natural Earth 1:50m boundaries via world-atlas; boundaries do not imply endorsement. Related reports: the geography of global vulnerability, Africa Ascendant and the polycrisis bill.