You walk toward a line. Science moves the line.
Think of the average end of life as a line ahead of you. Each year you live, you walk one year closer. Each year, medical progress pushes the line a little farther away.
If the line moves more slowly than you walk, you reach it in the end. If it moves as fast as you walk, you never do. That speed, one year of life expectancy gained for every year that passes, is what people call longevity escape velocity. The name borrows from the speed a rocket needs to break free of Earth's gravity.
Life expectancy is the average age people would reach if the death rates of one year stayed the same for their whole lives. CDC ↗PubMed Central ↗
At 3 months a year, an average 65-year-old would reach the line at about 91, 26 years from now.
In the CDC's 2024 tables a 65-year-old has 19.7 years left on average. The next birthday takes away 0.8 of a year. The gain from 2023 to 2024 gave back 0.2 of a year.
The lines are arithmetic to show the idea, with today's average remaining years as the starting gap. They aren't a forecast for anyone.
The careful version: how fast the risk of dying falls
The scientific form of the idea, in Aubrey de Grey's 2004 paper, is about risk. In the CDC's 2024 table, the chance of dying within a year is 1.3% at 65, 2.8% at 75, 7.8% at 85, and 22.5% at 95. It climbs about 10% with each year of age. To stay in place, the risk at each age would have to fall about that fast, every year. PubMed Central ↗CDC ↗
- Needed to stay in placeabout 10%
- United States in 2024, a rebound year3.8%
- The 12 months to September 20252.4%
- United States, average per year since 19790.82%
- Social Security's middle forecast to 21000.73%
How much the US death rate falls in a year, adjusted for age. CDC ↗CDC ↗Social Security ↗
The line has been crossed, but only by catching up
Each column is one real year. Almost all of them stay far below the line. The ones that reach it came right after a disaster, or while a country was catching up. None of these three has stayed above it for more than two years in a row.
+0.62024About 7 months gained in the year. Fewer overdose and COVID-19 deaths.
Source: CDC. Since 1950 the United States has crossed the line once: 2022, at +1.1, as deaths fell back after the worst of COVID-19, which had cost 2.4 years in 2020 and 2021.
See every year as numbers
The newest numbers, and the newest science
How current this is: US life expectancy estimated to September 2025 · US death rates to September 2025 · overdoses to June 2026 · official US life tables to 2024 · research counted this morning. The CDC publishes death rates about six to nine months after each quarter ends, and full life tables a year or more after. Sources last read Nov 21, 2026. New research takes years to show up in these numbers.
Some countries have held that pace for years
It has happened, and what happened is worth knowing. These countries gained more than a year per year for a decade or more. They were catching up with what others already had, mostly because far fewer children and young adults were dying. Then the pace fell away. Among the countries that live longest today, crossings since 1990 are short, and nearly all are rebounds after COVID-19. UN ↗
- 16 years in a rowBahrain, 1955 to 1970+22.7 years, catching up with richer countries
- 13 years in a rowOman, 1973 to 1985+16.6 years, catching up with richer countries
- 12 years in a rowRwanda, 1998 to 2009+19.8 years, after the 1994 genocide and during the HIV epidemic
- 8 years in a rowEswatini, 2012 to 2019+10.8 years, as HIV treatment spread
Months added per year
Taken over longer stretches, the rises and rebounds even out. From 2014 to 2024, US life expectancy went from 78.9 to 79.0: COVID-19 and overdoses took back almost everything gained in between. PubMed ↗PubMed ↗UN ↗CDC ↗
- 12Escape velocityWhat the idea needs
- 3The longest-lived countriesSince about 1840
- 2.4The world2013 to 2023
- 1.8United States1950 to 2024
- 1.6Japan2013 to 2023
- 0.1United States2014 to 2024
Overdoses and COVID-19 more than anything else
The CDC breaks down each year's change by cause of death. No treatment aimed at aging itself shows up on the list, because none has been approved. These numbers update here as the CDC publishes them.
What moved the US number in 2024
Life expectancy rose 0.6 years, from 78.4 to 79.0. The CDC splits that into what pushed it up and what held it back. CDC ↗
Share of the gains
- Fewer deaths from unintentional injuriesMostly overdoses: the overdose death rate fell 8.2 per 100,000, and the rate for all unintentional injuries fell 9.0.45.3%
- Fewer COVID-19 deaths10.8%
- Fewer heart disease deaths9.0%
- Fewer cancer deaths6.5%
- Fewer homicides5.9%
- All other causes22.5%
Share of what held it back
- More flu and pneumonia deaths41.4%
- More deaths from nutritional deficiencies33.0%
- More complications of medical care10.2%
- Other causes15.4%
deaths per 100,000, adjusted for age, in the 12 months to September 2025: 2.4% lower than a year earlier. CDC ↗
in the 12 months to June 2026, the CDC's estimate. That is 42% below the peak of 112,418 in August 2023. CDC ↗
The past year, cause by cause
Change in the US death rate over a year, per 100,000 people: the newest 12 months the CDC has for each cause against the same period a year before. Some causes take longer to count. CDC ↗
- Overdoses−6.8
- COVID-19−5.5
- Cancer−2.6
- Heart disease−2.0
- Homicide−0.9
- Falls, age 65 and over−0.9
- Alzheimer's−0.8
- Flu and pneumonia+1.4
Elsewhere, the latest year
An average is not a person
Life expectancy describes a crowd. People die in later life of specific illnesses, so a breakthrough against one illness lifts the crowd's average by saving the people who would have died of it. If that illness was never going to be yours, the breakthrough doesn't change the age you reach.
Critics of the idea say it treats a rising average as if everyone were gaining time. The CDC's own tables let you see what they mean.
One hundred people, placed at the age they die in the CDC's life table. The average is 76.9. It describes the crowd, and nobody in particular.
The dots use the CDC's 1999–2001 cause-of-death tables, the newest of this kind we could find, so each person's age and cause follow that table. The gains are the CDC's own. The tables remove deaths from a cause, not the illness itself. CDC ↗
It shows in the newest numbers too
Life expectancy at birth rose 0.6 years in 2024. For someone already 65 the rise in years left was 0.2, at 75 it was 0.1, and at 85 the table didn't move at all. Most of the headline gain came from fewer deaths before age 65. CDC ↗CDC ↗
- +0.6At birth78.4 to 79.0
- +0.5Age 2059.2 to 59.7
- +0.3Age 4040.7 to 41.0
- +0.2Age 6519.5 to 19.7
- +0.1Age 7512.4 to 12.5
- 0.0Age 856.7 to 6.7
- 0.0Age 904.6 to 4.6
Other doubts
The line can be crossed for the wrong reasons
US life expectancy jumped 1.1 years in 2022, which is over the line. It had just fallen 2.4 years in two years of COVID-19. A rebound is not progress against aging. CDC ↗
Each extra year gets harder
To add one more year in Japan in 2019, death rates at every age would have had to fall 20.3% for women and 9.5% for men, and that requirement has been growing since 1990. A 1990 paper in Science worked out that simply reaching an average of 85 would take a 55% fall in death rates at every age. PubMed Central ↗PubMed ↗
It is a snapshot, not a forecast
The CDC calls the life table a snapshot of one year's death rates. It says what would happen if those rates never changed. It doesn't predict how long anyone will live. CDC ↗
Escape isn't forever
Accidents and other causes don't stop. Someone who kept the risk of dying of a 25-year-old for life, 0.094% a year in the 2024 tables, would live about 1,060 years on average. That is our arithmetic. CDC ↗
Who would get it?
Life expectancy already differs by 14.7 years between groups in the US: 85.8 for non-Hispanic Asian people and 71.1 for non-Hispanic American Indian and Alaska Native people in 2024. A costly treatment could widen that. CDC ↗
What supporters answer
Aim at aging, not at one disease
Curing one disease at a time won't do it, supporters agree. Aging raises the risk of nearly all of them, so a treatment aimed at aging could help people whatever they would have died of. One simulation found that delayed aging would add 2.2 years by 2060, while separate progress on heart disease and cancer gave shrinking returns because other causes step in. PubMed ↗
It only has to buy time
De Grey's argument: a first treatment good enough to add some 20 healthy years gives researchers those 20 years to build a better one, and so on. PubMed Central ↗
Past ceilings were broken
Experts have repeatedly named a limit to life expectancy, and each was passed, on average about five years after it was published. Record life expectancy has risen steadily since about 1840. PubMed ↗PubMed ↗
We set out both sides and don't pick one. This site reports what has been measured and who has said what.
The forecasts say slowly. A few scientists say soon.
The lines are official forecasts, as published. The steep one is what escape velocity would have to look like. It is an illustration of the claim, not a forecast, and it leaves the chart within a generation.
Social Security Trustees, 2026
Middle forecast for the US: about 81.8 in 2056 and 85.3 in 2100. Their slower and faster cases reach 81.3 and 89.5 in 2100. Average of the published figures for men and women. Social Security ↗
Congressional Budget Office, 2026
For the US: 79.0 in 2026, rising to 82.3 in 2056. CBO ↗
United Nations, 2024
For the US: 83.2 in 2050 and 89.2 in 2100. For the world: 77.0 in 2050 and 81.7 in 2100. The UN's US figures start higher than the CDC's because it estimates them differently. UN ↗
The Lancet, 2024
For the world: from 73.6 in 2022 to about 78.1 in 2050, a slower rise than in the 30 years before COVID-19. PubMed ↗
Who says what, and when
- 2029–2035
Ray Kurzweil Inventor and author
Writing in The Economist in 2024, he put escape velocity between 2029 and 2035, helped by AI that can simulate biology. In 2018 he had said about 2028 to 2030. The Economist ↗Wikipedia ↗
- By 2036
Aubrey de Grey Biologist, LEV Foundation
He has given it a 50% chance by 2036. He argues the first treatments need not be perfect, only good enough to buy time for better ones. Wikipedia ↗PubMed Central ↗
- About 2050
George Church Harvard geneticist
He said he wouldn't be surprised if 2050 were the point, and stressed that such estimates are very uncertain. Interview ↗
- This century
James Vaupel and colleagues Demographers
If today's steady pace simply continues, most children born this millennium will reach 100. That is about 3 months a year, not 12, and they say forecasts could prove too high or too low. PubMed ↗
- Not this century
S. Jay Olshansky and colleagues Demographers
Gains have slowed since 1990. Unless biological aging can be markedly slowed, they consider radical life extension implausible this century, and expect at most about 15% of women and 5% of men to reach 100. PubMed Central ↗
Treatments aimed at aging itself
Official forecasts assume today's slow progress carries on. Going much faster would take treatments that slow aging itself. Here is how far the best-known ones have got. A research stage shows how far testing has gone, not whether something works.
- Cell reprogrammingStage 2 of 4 · Small studies in peopleThe first treatment of this kind, ER-100, went into people in 2026, in the eye. First company-reported results cover 3 people at the lowest dose.
- RapamycinStage 2 of 4 · Small studies in peopleExtends the lives of mice. Studies in people are small so far.
- MetforminStage 4 of 4 · Widely recommendedIts high stage comes from its approval for diabetes, not aging. The planned TAME trial would test whether it slows aging itself. As of October 2026 that trial hasn't started: money is still being raised.
- SenolyticsStage 2 of 4 · Small studies in peopleDrugs meant to clear out damaged, aged cells. Trials in people are small so far.
- GLP-1 drugsStage 4 of 4 · Widely recommendedApproved for diabetes and obesity, not for aging. Researchers are testing whether they do more than that.
Can aging be reversed? → Every treatment by research stage →
New research on how long we live
- Eating more kinds of food may link to slower early frailty in older Chinese adults
- Long-lived tortoise and human centenarians show similar DNA patterns
- How the hearts of people aged 100 and older hold up over time
- Frailty linked to fewer healthy working years after age 50
- Exercise and social mental activity linked to more years free of memory decline
Escape velocity, plainly
What is longevity escape velocity?
It's the idea that if life expectancy rises by more than one year for every year that passes, people would stop running out of time, because remaining life expectancy would grow faster than they use it up. The name comes from the speed a rocket needs to break free of Earth's gravity. It is also called actuarial escape velocity. PubMed Central ↗Wikipedia ↗
Who came up with it?
The idea was around in life-extension circles by the 1970s. The biologist Aubrey de Grey gave it scientific form in a 2004 paper, and credit for the term goes to David Gobel of the Methuselah Foundation. The inventor Ray Kurzweil made it famous. PubMed Central ↗Wikipedia ↗
Has any country reached it?
For stretches, but not in the way the idea means. Rwanda gained more than a year per year for 12 years in a row from 1998, and Bahrain for 16 years from 1955, while catching up with the rest of the world, mostly because far fewer children and young adults died. Among the countries that live longest today, the crossings since 1990 are short, and nearly all are rebounds after COVID-19. The United States has crossed the line once since 1950, in 2022, and Japan six times between 1951 and 1963. CDC ↗UN ↗
How fast is life expectancy actually rising?
In the countries that live longest, about 3 months a year on average since about 1840. Gains slowed after 1990 in most long-lived countries. In the United States, life expectancy was 78.9 in 2014 and 79.0 in 2024, because of COVID-19 and overdoses in between. PubMed ↗PubMed Central ↗CDC ↗
If life expectancy goes up, will I live longer?
Not necessarily. Life expectancy is an average for a whole population, worked out from one year's death rates. It can rise because fewer younger people die of overdoses, which changes little for someone who is 85. In 2024 it rose 0.6 years at birth and didn't change at all at age 85. CDC ↗CDC ↗
Would reaching escape velocity mean living forever?
No. Accidents and other causes of death would remain. Someone who kept a 25-year-old's risk of dying for life would live about 1,060 years on average, using the 2024 US life tables. CDC ↗
When do supporters expect it?
Ray Kurzweil has said between 2029 and 2035, Aubrey de Grey has given it a 50% chance by 2036, and George Church said he wouldn't be surprised by 2050. Official forecasts expect far slower gains: Social Security's middle forecast has US life expectancy at about 81.8 in 2056. The Economist ↗Wikipedia ↗Interview ↗Social Security ↗
Is there a treatment that does this now?
No. No treatment has been approved for slowing aging itself, and none has been proven to lengthen human life by slowing aging. Several are being tested. Talk to your doctor before starting or changing any treatment. PubMed Central ↗
How this page stays current
The death-rate, overdose, European and United Nations numbers refresh by themselves each morning when those sources publish something new; they were last read Nov 21, 2026. New annual US figures come in from the CDC's yearly report. Because those yearly figures arrive a year or more late, the dial and the “Right now” panel use our own estimate of US life expectancy for the newest 12 months of the CDC's provisional death rates by age, lined up with the newest official year, and count the research on aging itself that the site's daily update found in the last 30 days. Everything else on this page was checked by hand on Nov 7, 2026. Provisional CDC numbers can change. Nothing here is our own forecast, and the things we worked out ourselves (the newest-12-months estimate, the what-if with the 100 people and the 1,060-year sum) say so where they appear.
- NCHS, United States Life Tables, 2024 (National Vital Statistics Reports 75-5, August 25, 2026) ↗
- NCHS, United States Life Tables, 2023 (National Vital Statistics Reports 74-6) ↗
- NCHS Data Brief 548, Mortality in the United States, 2024 (January 2026) ↗
- NCHS, United States Life Tables Eliminating Certain Causes of Death, 1999–2001 (National Vital Statistics Reports 61-9, 2013) ↗
- CDC, quarterly provisional estimates for selected indicators of mortality (dataset 489q-934x) ↗
- CDC, provisional drug overdose death counts (dataset xkb8-kh2a) ↗
- United Nations, World Population Prospects 2024 ↗
- Our World in Data, life expectancy (Human Mortality Database and UN) ↗
- 2026 Social Security Trustees Report, Table V.A4: period life expectancy ↗
- 2026 Social Security Trustees Report, demographic assumptions ↗
- Congressional Budget Office, The Demographic Outlook: 2026 to 2056 ↗
- GBD 2021 Forecasting Collaborators, The Lancet, 2024 ↗
- Olshansky and others, Implausibility of radical life extension in humans in the twenty-first century, Nature Aging, 2024 ↗
- Olshansky, Carnes and Cassel, In search of Methuselah, Science, 1990 ↗
- Oeppen and Vaupel, Broken limits to life expectancy, Science, 2002 ↗
- Vaupel, Villavicencio and Bergeron-Boucher, Demographic perspectives on the rise of longevity, PNAS, 2021 ↗
- Goldman and others, Substantial health and economic returns from delayed aging, Health Affairs, 2013 ↗
- de Grey, Escape velocity: why the prospect of extreme human life extension matters now, PLoS Biology, 2004 ↗
- Aubrey de Grey, post on X, as quoted by Wikipedia ↗
- Ray Kurzweil, The Economist, June 17, 2024, as summarized by Wikipedia ↗
- George Church, interview with Dwarkesh Patel, August 2025 ↗
- Wikipedia, Longevity escape velocity ↗
- Japan Ministry of Health, Labour and Welfare, abridged life table, 2025 ↗
- Eurostat, life expectancy by age and sex (demo_mlexpec) ↗
This page is general information, not medical advice. Talk to your doctor before you start, stop or change any treatment.

