Key figures (as of 2 November 2026)
| Undernourished people, 2025 (July 2026 report, point estimate) | 644.6 million (headline 645 million), 7.8% of the world population |
| Uncertainty range, 2025 (July 2026 report) | 608 million to 696 million (7.4% to 8.5%) |
| Undernourished people, 2024, as revised in the July 2026 report | 658.8 million (8.1%) |
| Undernourished people, 2024, as first published (July 2025 report) | 673.2 million (headline 673 million, range 638 million to 720 million) |
| Undernourished people, 2022 to 2024, July 2026 report | 687.6 (2022), 686.4 (2023), 658.8 (2024) million |
| Projection for 2030 (July 2026 report) | 510 million to 520 million; scenarios of 503, 511 and 522 million |
| Africa, 2025 (July 2026 report) | About 309 million people, 20.0% of the population, down from a peak of 20.3% in 2024 |
| Asia, 2025 (July 2026 report) | About 292 million people, 6.0% of the population |
| Latin America and the Caribbean, 2025 (July 2026 report) | About 32 million people, 4.8% of the population |
Three years of gradual progress
The July 2026 report puts the number of undernourished people at 644.6 million in 2025, down from 658.8 million in 2024 and 687.6 million in 2022. The share of the world population facing hunger fell from 8.6% in 2022 to 8.1% in 2024 and 7.8% in 2025. The number is still above its 2019 level of 583.7 million, before the sharp rise in 2020 and 2021.
Progress has been uneven. In Asia, the number fell from 364.3 million in 2022 to 291.8 million in 2025, led by Southern Asia and South-eastern Asia. In Latin America and the Caribbean it fell to about 32 million. In Africa the share fell slightly in 2025, the first break in the rise since 2017, but the number still edged up, to 309.2 million, because the population is growing. In Western Asia the share rose to 13.0% in 2025. The report notes that the countries facing the deepest humanitarian crises are also those where recent data are often missing, so estimates of chronic hunger there may be less reliable.
First-published figures and revisions
Each edition revises the whole series, as countries send new data on food supply, as new household surveys arrive, and as population estimates are updated. The report itself urges readers not to compare figures from different editions. The figure for the latest year is always a nowcast, and it has been revised in the next edition each time:
| Year | First published | Revised in the next edition | Change (Centre's calculation) |
|---|---|---|---|
| 2022 | 735.1 million (2023 edition) | 723.8 million (2024 edition) | -11.3 million |
| 2023 | 733.4 million (2024 edition) | 688.4 million (July 2025 report) | -45.0 million |
| 2024 | 673.2 million (July 2025 report) | 658.8 million (July 2026 report) | -14.4 million |
All three revisions were downward, and the largest was for 2023. The July 2025 report attributes much of the improvement in Southern Asia to new data from India. This study counts only the first-published figure for 2026, so later revisions do not change the result.
The first-published point estimate has fallen in each of the last three editions: from 735.1 million (2022) to 733.4 million (2023), 673.2 million (2024) and 644.6 million (2025). In each edition, the change from the previous year to the nowcast year was -3.8 million (2023 edition), +9.6 million (2024 edition), -15.2 million (July 2025 report) and -14.2 million (July 2026 report), on the Centre's own calculation.
The outlook for 2026
The July 2026 report sets out three projection paths to 2030:
- Before the 2026 disruptions: the number falls to 503 million by 2030.
- Quick recovery: energy and fertilizer supplies recover quickly and fully in the second half of 2026; the number falls to 511 million by 2030. The report calls this a lower-bound estimate of the effect of the disruptions.
- Lasting crisis: higher energy and fertilizer costs persist and recovery takes until 2030; the number falls to 522 million by 2030.
On the Centre's own calculation, a straight path from 645 million in 2025 to these 2030 levels means a fall of about 25 million to 28 million a year. The report's scenarios focus on structural trends, and the shock fell in 2026 itself, so the Centre expects a smaller fall in 2026 than this average. The report also says that weather extremes, such as El Niño in 2026 and 2027, are not included in these estimates and could add to the negative effects, and the launch release notes that recent reductions in development and humanitarian financing are not fully accounted for either.
What it means
The world looks set for a fourth year of falling hunger on the first-published figure, but the size of the fall depends on revisions as much as on events in 2026. On the Centre's estimates, the combined weight on a figure below 635 million is 67%, and the combined weight on a figure below 650 million is 85%. The combined weight on 650 million or more, a rise from the 2025 nowcast, is 15%.
Scenarios: the number of undernourished people in the world in 2026
These probabilities are the Centre's analytical estimates, reasoned from the figures above. Because the bands are ordered, the forecast is scored with both the multi-category Brier score and the ranked probability score.
The variable is the number of undernourished people in the world in 2026, in millions: the world point estimate (not the lower or upper end of the range) given in the annual joint UN report on world food security and nutrition, as first published in the edition that first covers 2026, expected in July 2027. Where the report gives the figure to one decimal place, as in its table of the number of undernourished people, that figure is used; it is rounded once to the nearest whole million, half up, so 619.5 million counts as 620 million. Only the first-published figure counts: later revisions do not change the result. The bands are half-open intervals and each edge falls in the higher band: exactly 590 million falls in band 2, exactly 605 million in band 3, exactly 620 million in band 4, exactly 635 million in band 5, exactly 650 million in band 6 and exactly 665 million in band 7.
If that report has not published a 2026 figure by 31 December 2027, the study resolves on the single-year (annual value) 2026 value of the world number of undernourished people in the official global statistics database of the UN food agency as shown on 31 December 2027, rounded the same way. The database's three-year average never resolves the study. If the report or the agency is renamed, merged or replaced, the study resolves on the successor's first-published world figure for 2026 on the same definition. If neither the report figure nor the database value exists by 31 March 2028, the study is withdrawn. The last date for any dated revision of these probabilities is 30 June 2027, before the expected July 2027 publication; no revision is accepted after that.
How the Centre builds its estimate. The Centre's starting point is the 2025 nowcast of 644.6 million. Two adjustments follow, both on the Centre's own calculation. First, the 2025 figure is likely to be revised: the last three revisions averaged about -24 million, but only about -13 million without the unusual 2023 case. Second, the change from 2025 to 2026: the last two editions showed falls of about 14 million to 15 million in the nowcast year, and the projection paths average 25 million to 28 million a year, but the 2026 disruptions point to a smaller fall in 2026. Together these point to a figure of about 615 million to 630 million, with wide uncertainty on both sides.
Sharp fall: below 590 million
This band needs a fall of more than 54 million from the 2025 nowcast. It would take another very large downward revision, like the 2023 case, together with continued progress in 2026. The weight is low.
Strong fall: 590 million to below 605 million
This band means a fall of about 40 million to 55 million from the 2025 nowcast. It would follow from a downward revision larger than usual plus a fall in 2026 close to the projection paths. The weight is modest.
Steady progress: 605 million to below 620 million
This band covers a fall of about 25 million to 40 million. It matches a typical downward revision plus a 2026 fall close to those of the last two editions. The weight is solid.
Slower progress: 620 million to below 635 million
This band covers a fall of about 10 million to 25 million. It fits a modest revision and a smaller fall in 2026 as the energy and fertilizer shock slows progress. It is the most likely band.
Near standstill: 635 million to below 650 million
This band covers a figure from about 10 million below the 2025 nowcast to about 5 million above it. It would follow if revisions are small and the 2026 disruptions, weather extremes or lower aid stall progress. The weight is solid.
Rise: 650 million to below 665 million
This band means a rise of about 5 million to 20 million from the 2025 nowcast. It would need an upward revision or a clear setback in 2026. The weight is modest.
Sharp rise: 665 million or more
This band means a rise of about 20 million or more. It would take a serious global setback, such as a lasting energy and fertilizer shock combined with a strong El Niño, or an upward revision of a kind not seen in the last three editions. The weight is low.
What to watch
- The July 2027 report: the first-published 2026 figure, and how far it revises 2025.
- Energy and fertilizer costs: whether supplies recover quickly from the 2026 disruptions or stay tight into 2027.
- El Niño: its effect on harvests in 2026 and 2027, which the projections do not include.
- Humanitarian and development financing: whether recent reductions continue.
- Africa and Western Asia: whether the first fall in Africa's share holds, and whether the rise in Western Asia continues.
- New data: updated food supply data and large household surveys, which together drove the largest recent revision.
Figures as of 2 November 2026. Probabilities are analytical estimates by the Ghaf Centre for Studies & Research.