India's Air Pollution, Report by Report: What the Big 2025–26 Studies Say
IQAir, the University of Chicago, the Lancet, State of Global Air, CREA and CSE all measure India's pollution differently. Here is what each found, why the numbers don't match, and what the courts and the law are doing about it.
India has no shortage of pollution reports. It has a shortage of one agreed number. Depending on whether a study counts micrograms in the air, deaths, years of life or rupees spent, India's air problem is described in very different ways. That is why headlines this year have said "sixth most polluted country", "two million deaths" and "3.5 years of life lost", often about the same air.
This article sets the main reports side by side. It says what each one measured, who published it, and where the evidence is firm and where it is not. It then covers the law and the courts, and what it means for a business operating in India.
The reports at a glance
| Report | Published by | What it measures | India headline | Data year |
|---|---|---|---|---|
| World Air Quality Report 2025 | IQAir | Annual average PM2.5 in the air | 48.9 µg/m³, 6th highest country | 2025 |
| Air Quality Life Index (2025 update) | Energy Policy Institute, University of Chicago | Life expectancy lost to PM2.5 | 3.5 years lost on average | 2023 |
| State of Global Air 2025 | Health Effects Institute and IHME | Deaths from air pollution | About 2 million deaths | 2023 |
| Lancet Countdown 2025 | Lancet Countdown | Deaths and economic value from PM2.5 | About 1.72 million deaths | 2022 |
| Ten-city mortality study | The Lancet Planetary Health | Daily deaths linked to PM2.5 | 7.2% of deaths in ten cities | 2008–2019 |
| Tracing the Hazy Air 2026 | CREA | Progress of the National Clean Air Programme | 77 of 100 cities cut PM10 | Review of Jan 2026 |
| State of India's Environment 2026 | CSE and Down To Earth | Whether monitoring keeps up | Over 64% of districts lack continuous monitoring | 2026 |
The first four rows are the headline numbers. They differ because they count different things, not because some are wrong. Keep that in mind throughout.
What the international reports say
IQAir, concentration. IQAir's 2025 report covers 9,446 cities in 143 countries, regions and territories. It puts India's average at 48.9 µg/m³ of PM2.5, sixth in the world behind Pakistan (67.3), Bangladesh (66.1), Tajikistan, Chad and the Democratic Republic of the Congo. The WHO's annual guideline is 5 µg/m³, so India's average is close to ten times that. On cities, Loni (Ghaziabad district) is the most polluted in the world at 112.5 µg/m³. Byrnihat in Meghalaya is third (101.1), Delhi fourth (99.6) and Ghaziabad seventh (89.2). Of the world's 50 most polluted cities, 29 are in India. Only 14% of the cities worldwide met the WHO guideline.
The University of Chicago's AQLI, life expectancy. The AQLI turns satellite-based PM2.5 estimates into years of life. For 2023 it puts India's population-weighted average at 41 µg/m³. Meeting the WHO guideline would add 3.5 years to the average Indian's life. All 1.4 billion people live above that guideline, and 46% live above India's own annual standard of 40 µg/m³. Delhi's residents would gain 8.2 years, and the seven states and territories of the Northern Plains, with 544.4 million people, would gain about 5. For comparison, the AQLI puts child and maternal malnutrition at 1.6 years of life lost and tobacco at 1.5. It also notes that 2023 was worse than 2022, which had been a record year of decline.
State of Global Air, deaths. The 2025 edition, from the Health Effects Institute and the Institute for Health Metrics and Evaluation, estimates about 2 million air-pollution deaths in India in 2023. That is up 43% from 1.4 million in 2000. About 89% were from non-communicable diseases such as COPD, heart disease, lung cancer, diabetes and, for the first time in the series, dementia. Maharashtra, Uttar Pradesh, Rajasthan, Bihar and West Bengal each had over 100,000 deaths. Deaths from household smoke fell, while deaths from outdoor PM2.5 and ozone rose.
Lancet Countdown, deaths and money. Released on 29 October 2025, it attributes about 1.72 million deaths in India in 2022 to human-caused PM2.5, up 38% since 2010. It values those premature deaths at roughly US$339 billion, equal to about 9.5% of GDP. Down To Earth describes this as the monetised value of premature deaths, not a measured fall in output, and its coverage is our source for these numbers.
The Lancet Planetary Health, ten cities. A 2024 study by researchers from Karolinska Institutet, Ashoka University and others examined ten cities from 2008 to 2019. It linked 7.2% of all deaths there, about 33,000 a year, to PM2.5 above the WHO daily guideline. Delhi was highest at 11.5% and Shimla lowest at 3.7%. The WHO daily guideline was exceeded on 99.8% of days. Mortality stayed high even on days below India's national limit of 60 µg/m³.
What the Indian reports say
CSE, the monitoring gap. The State of India's Environment 2026, released on 25 February 2026, asks a prior question: can India see its own air? CSE counts 562 real-time monitors across 294 cities and 966 manual stations across 419 cities and towns. Over 64% of India's 742 districts have no continuous monitor. Only about 15% of the population, around 200 million people, live within 10 km of one. The satellite-based and modelled estimates above fill that gap, which is one reason the reports differ.
CREA, the clean air programme. The National Clean Air Programme began in 2019 with a target of cutting particulate pollution 20–30% against 2017 levels by 2024. In 2022 the target was raised to 40% by 2026 for an expanded list of cities. CREA's January 2026 review looked at 100 cities with enough PM10 data:
- 77 cities reduced PM10 and 23 saw it rise.
- Of the 77, 68 were still above the national standard.
- 51 cities met the original target and 23 met the revised 40% target.
- ₹13,415 crore had been released and ₹9,929 crore (74%) used.
- About 68% of spending went to road dust, 14% to transport and 12% to waste and biomass burning. Industry, domestic fuel and public outreach each received under 1%.
CREA recommends funding cities according to studies of where their pollution comes from, putting PM2.5 ahead of PM10, and planning by regional "airshed" as well as by city.
CPCB, rivers. Pollution is not only in the air. The Central Pollution Control Board's 2025 assessment of polluted river stretches found 296, down from 311 in the previous assessment. Locations unfit for bathing fell from 815 to 807. Maharashtra has the most polluted stretches (54), then Kerala (31), and Madhya Pradesh and Manipur (18 each). The most urgent category, with BOD above 30 mg/L, has five stretches each in Tamil Nadu, Uttar Pradesh and Uttarakhand. We have read this through press coverage, not the CPCB document itself.
Why the numbers don't match
- Different things are counted. IQAir measures air. The AQLI measures life. State of Global Air counts deaths from outdoor, household and ozone pollution; Lancet Countdown counts deaths from human-caused PM2.5. You cannot put them in one sentence.
- Different years. The data runs from 2022 to 2025. Pollution moved between them.
- Deaths are modelled, not recorded. No death certificate says "air pollution". Every death figure above is a statistical estimate. In December 2025, the Union Health and Environment ministries told the Rajya Sabha, as reported, that there is no conclusive data to establish a direct link between deaths exclusively and air pollution. Opposition MPs called the replies misleading. The studies themselves are statistical estimates of how many deaths exposure contributes to, not counts taken from death certificates.
- Coverage is uneven. City rankings depend on where monitors are, which is exactly the gap CSE describes.
What all of them agree on is the direction: India's air is above the WHO guideline everywhere, above its own standard for nearly half the population, and worst in the Northern Plains.
The law and the courts
The legal framework is the Air (Prevention and Control of Pollution) Act, 1981, the Environment (Protection) Act, 1986, and the National Green Tribunal Act, 2010. The Supreme Court has read the right to a pollution-free environment into the right to life under Article 21. Subhash Kumar v. State of Bihar (1991) is the case usually cited for it. The M.C. Mehta litigation has shaped Delhi's air policy over decades. Down To Earth credits it with Delhi's shift of public transport to CNG, the creation of the Environment Pollution (Prevention and Control) Authority, the phase-out of older vehicles and the Graded Response Action Plan (GRAP) for winter smog.
The courts are still pressing. On 6 January 2026, a bench of Chief Justice Surya Kant and Justice Joymalya Bagchi said the status note of the Commission for Air Quality Management (CAQM) seemed to ignore most of the issues raised and that the Commission seemed in no hurry to identify causes or long-term solutions. It directed CAQM to meet domain experts, put the major causes on the public record and begin phased, long-term measures.
The winter is already here. On 3 November 2026 Delhi's AQI was 350, "very poor" and the worst since 20 January. CAQM had invoked Stage II of GRAP the day before, when the AQI hit 324. Satellite detections showed 3,535 farm fires from 15 September to 2 November, 219 of them on 2 November alone.
What it means for a business
This is our reading, not legal advice.
- Winter curbs are an operating risk. GRAP stages bring temporary curbs on activities that raise dust and emissions whenever the air turns bad. If you operate in Delhi-NCR, plan for them as you would for a monsoon.
- Expect the focus to move. CREA's spending data shows industry and domestic fuel barely funded, and the Supreme Court is asking for sources to be identified. Our reading is that rules will become more source-specific over time, so industrial and transport operators should expect closer attention.
- The data will improve. CSE's monitoring gap is now a public talking point. Better coverage means more evidence available to regulators, courts and investors, so keep your own emissions and compliance records in order.
- Reports are not law. None of these studies binds anyone. They matter because courts, regulators and customers read them, and because they set the numbers that policy is argued over.
What this article does not cover
- Not covered: soil, plastic and noise pollution, and any change to penalty provisions in environmental statutes.
- Not verified at source: we read the AQLI fact sheet, CREA's publication page, IQAir's rankings and the Karolinska release directly. We did not open the State of Global Air 2025 or Lancet Countdown 2025 reports, the CSE report itself, or the CPCB river report. Their figures come from the press coverage linked below, so check them against the originals before relying on them.
- Not found: a 2026 update of the AQLI and a State of Global Air 2026 edition, as of 9 November 2026. IQAir's count of Indian cities among its 100 most polluted is also not verified; we quote only the top 50.
Sources
- IQAir: World Air Quality Report 2025 · most polluted countries · most polluted cities
- University of Chicago, EPIC: AQLI India fact sheet (2025 update)
- Health Effects Institute: State of Global Air · Down To Earth: coverage of SoGA 2025
- Down To Earth: coverage of Lancet Countdown 2025
- Karolinska Institutet: ten-city mortality study
- CREA: Tracing the Hazy Air 2026
- Down To Earth: CSE's State of India's Environment 2026
- Drishti IAS: CPCB polluted river stretches 2025
- Down To Earth: Supreme Court order of 6 January 2026 · the M.C. Mehta litigation
- The Week (PTI): Delhi air quality, 3 November 2026 · Government's reply on pollution deaths, December 2025
- Down To Earth: WHO 2021 air quality guidelines