To main content To menu
Chirps About Contact

The Brief Spike 8 billion people and falling: a brief spike in human history

Global population peaks before 2080, then falls. Four-fifths of all human births are already behind us. What that means for the next two centuries.
EN· ·Published inCuriosity & Science
8 billion people and falling: a brief spike in human history

On November 15, 2022, we crossed 8 billion humans. Today, in May 2026, the global population is estimated to be around ≈8.3 billion. Depending on which forecaster you trust, our population will peak at between ≈9.7 and ≈10.4 billion, after which it will start to fall, possibly forever. Cyberpunk lied to us. There won’t be a trillion-person megacity Earth.

Let me run the numbers:

The three main forecasts disagree, but only on the details

The three institutions that conduct this type of research are the United Nations Population Division (WPP), the International Institute for Applied Systems Analysis (IIASA) in Vienna, and the Institute for Health Metrics and Evaluation (IHME) at the University of Washington. Their latest medium-variant projections of the global population peak diverge in interesting ways:

For comparison, the world population was 1 billion in 1800, 2.3 billion in 1940, 3.7 billion in 1970, and 7.4 billion in 2016. The annual growth rate peaked at 2.2% in 1964 and has been falling ever since, reaching ≈0.9% in 2023 and projected to hit -0.1% by 2100.

The story isn’t really about whether we hit 9.7 or 10.4 billion. It’s about the fact that the global total fertility rate (TFR) – the average number of children a woman has in her lifetime – has dropped from ≈5 in 1950 to ≈2.25 in 2024, with two-thirds of the world’s population living in countries with sub-replacement fertility (i.e. below ≈2.1). The UN expects the global TFR to fall below replacement level around 2054. After that, momentum will carry us up for another generation, and then mathematics will take over.

The spike

The part that broke my brain is this: in April 2024, Spears, Vyas, Weston, and Geruso published a paper in PLOS ONE that extended the projections beyond 2100 – something that had not been done since 2013. Their key finding is quoted verbatim from the abstract below because it deserves to be precise:

Without such an increase [back to replacement fertility], the 400-year span when more than 2 billion people were alive would be a brief spike in history. Indeed, four-fifths of all births – past, present, and future – would have already happened.

Read that again. The Population Reference Bureau estimates that ≈117 billion humans have ever been born. If global fertility stays below replacement (and right now it’s heading there fast), only ≈26 billion more people will be born. Total ceiling: ≈150 billion. Statistically, we are in the final 20% of human births that will ever occur.

At a sustained TFR of 1.2 (roughly the rate in East Asia today), Spears et al. show the world population would halve every ≈40 years. By 2280, annual births would be down to ≈10 million globally – vs ≈132 million today. By 2340, the global population would fall below 1 billion again. This isn’t a forecast but rather a conditional projection: if fertility rates remain as they are, this is what the numbers would look like. If fertility recovers, WHERE we stabilize depends on WHEN it recovers. A return to replacement level fertility starting in 2125 would stabilize the world population at ≈8 billion. Starting in 2175 would stabilize us at ≈2-4 billion. Starting in 2200 with a prior TFR of 1.5 stabilizes us at ≈3 billion.

Timing matters more than rate.

Why fertility falls and refuses to come back up

The structural drivers are reasonably well understood: access to contraception, female education, the time required to establish oneself in the labor market, housing costs, the opportunity cost of women’s careers, and the rising age at first birth (the EU average is now 31.1 years). What’s contested is whether a country can climb back out once it falls deep enough.

Lutz, Skirbekk, and Testa (2006) termed this the low fertility trap hypothesis. Below a total fertility rate (TFR) of ≈1.5, three reinforcing mechanisms kick in:

  1. Demographic momentum
    Fewer potential mothers in the next generation.
  2. Sociological adaptation
    Younger cohorts grow up around smaller families and internalize lower ideal family sizes.
  3. Economic squeeze
    Young people’s expected incomes fall (partly due to aging-related fiscal pressure) while their aspirations rise.

The empirical evidence is uncomfortable: South Korea spent ≈270 billion USD on pro-natalist incentives over 16 years. The TFR decreased from 1.21 in 2010 to a record low of 0.72 in 2023. However, Statistics Korea reported a slight increase to 0.75 for 2024 in early 2025 and to 0.80 for 2025 in early 2026 – the first sustained uptick in nine years – largely due to a post-Covid marriage backlog. Yet most demographers expect this increase to be temporary; the 30-something cohort driving it will age out, and Seoul’s TFR remains at ≈0.64, likely the lowest of any geographic unit on Earth.

Arguably the most aggressive pro-natalist regime on the planet, Singapore sits at a TFR of 0.87 in 2025. Hungary, Poland, France, and Scandinavia have experienced minimal sustained impact from cash incentives, parental leave, free preschool, and so on. The general lesson is that money doesn’t buy babies. Or rather, it doesn’t buy enough of them.

Two countries, two trajectories: Vietnam and Germany

Closer to home, the same dynamic is playing out at very different speeds: Vietnam crossed below replacement around 2023. According to the General Statistics Office, the total fertility rate (TFR) was 2.11 in 2021, 2.01 in 2022, 1.96 in 2023, and 1.91 in 2024. The latest figure for 2025 shows a slight uptick, but only ≈11 of the 63 provinces are below 2 children per woman, while 19 are still above 2.2 – meaning the urban south (Sài Gòn, Mekong Delta TFR ≈1.48-1.62) is already well below replacement level, while the northern midlands are at ≈2.34. Vietnam’s population peak is projected to be around ≈110 million in 2050. The demographic dividend window will close around 2039-2040.

Vietnam’s new Population Law, passed in December 2025 and effective July 1, 2026, eliminated the two-child limit and introduced incentives, including extended maternity leave, social housing priority for families with two or more kids, and longer paternity leave. After decades, the phrase “đẻ ít con để nuôi dạy cho tốt” (“have few children to raise them well”) is officially gone. The unspoken concern: Vietnam might age before it becomes wealthy. The sex ratio at birth was 111.4 boys per 100 girls in 2024, which is a structural problem inherited from the same era of family planning that is now being undone.

Germany is one demographic generation ahead. Destatis reports that the TFR fell from 1.38 in 2023 to 1.35 in 2024. Births in 2024 were ≈677,000. In 2025, the birth count was ≈3.4% lower – the lowest birth count since 1946. Germany has had more deaths than births every single year since 1972 – 53 consecutive years of natural decline, masked entirely by immigration. In 2025, the share of the population with a foreign background reached 31.1%. The mean maternal age at first birth is now 31.8 years, up 3.9 years since 1991. Berlin’s total fertility rate (1.21) is now lower than the former East German average (also ≈1.21). The East-West fertility gap, which used to exist, has flipped and collapsed.

“But what about Africa?”

This is the question that always comes up. Sub-Saharan Africa is truly the demographic engine of the rest of this century – it is currently home to ≈1.2 billion people and is projected to grow to over ≈3.4 billion by 2100 according to the UN medium variant. Sub-Saharan Africa’s share of global births will grow from ≈29% in 2021 to ≈54% in 2100, based on IHME projections.

But here’s the catch: SSA’s population growth is slower than the UN’s forecasts predicted. In the UN’s 2012 forecast, Nigeria’s projected population in 2100 was 914 million. The 2022 forecast was 546 million. And the latest 2024 forecast is 477 million. That’s a downward revision of 437 million in just 12 years – more than the current population of the United States, gone from a single country’s projection. IHME projects that Niger’s total fertility rate (TFR), currently the world’s highest at ≈6.5, will fall below replacement level by the 2080s. The Club of Rome’s 2024 Earth4All update suggests that Sub-Saharan Africa could peak as early as 2060 – 40 years earlier than the UN’s medium variant.

Female education and urbanization are achieving in two generations what took four in Europe. The trajectory could bend upward in extreme scenarios – the UN’s SSP3 “regional rivalry” path projects 12.6 billion by 2100. But the central tendency is downward acceleration.

What will a civilization built on growth do when it stops growing?

This is where things get philosophically uncomfortable: for roughly 250 years – since the Industrial Revolution – the entire economic, political, and ideological architecture of modern civilization has been based on the assumption of growth. Not just economic growth, but also demographic growth: more workers next year than this year. More taxpayers funding more pensioners. More consumers buying more stuff. More renters bidding up housing prices. More soldiers will defend more territory. Every single major institution we’ve inherited – capitalism, social security systems, public debt, real estate as wealth, and even the modern nation-state – assumes the population pyramid will remain roughly pyramid-shaped.

It doesn’t anymore. And the inversion is happening quickly enough that some of us will live through both ends of it.

Pensions and public debt will go from “demanding” to structurally impossible

Pay-as-you-go pension systems, such as those in Germany, France, Japan, Korea, and the US, are mathematically a wager that future workers will outnumber future retirees. Spears et al’s TFR 1.2 long-run scenario produces an old-age dependency ratio of ≈150 – more than one retiree per working-age adult. Today, Japan, the world’s oldest society, already sits at ≈50. There is no historical precedent for a society with an old-age dependency ratio (OADR) of 150, and no version of “raise the retirement age” or “raise contributions” can close that gap. Either pensions will become unrecognizable (the optimistic version is a universal basic income from AI productivity), or they will collapse.

Real estate stops being a one-way bet

Property, the global middle-class’s preferred wealth-storage instrument, only works when more people want housing next year than this year. Japan currently has 9 million vacant homes – 13.8% of all housing stock – and Nomura projects that figure will rise to ≈30% by 2033. The Japanese word for these homes, akiya (空き家), is becoming a cross-cultural import. Apartments being built in South Korea today are projected to enter long-term vacancy by the 2050s. China’s “ghost city” phenomenon – infrastructure built based on unrealized growth projections – is already becoming a permanent feature rather than a temporary mismatch in places like Ordos and parts of Tianjin.

The land never goes empty. The buildings do. And the financial system that priced these buildings as appreciating assets must absorb that reality.

Schools, hospitals, infrastructure: the receding tide

Since 1980, South Korea has shut down 4,008 schools, as student enrollment fell from 9.9 million to 5.07 million. 376 of those schools are still standing, unused. Ghost schools. 82 have been abandoned for over 30 years. Between 2018 and 2022, the number of pediatric clinics in Seoul fell by 12.5%, partly because there aren’t enough children to justify them economically. Meanwhile, the number of psychiatry clinics rose by 76.8%, and the number of anesthesiology centers by 41.2%. Medical specialties are shifting toward the elderly because that’s where the patients are. Subway stations get fewer riders, then close. Bus routes consolidate. Then, they’re canceled. This process is happening across every aging society, resulting in a slow, geographically uneven hollowing out of the 20th-century physical infrastructure.

Geopolitics, the military, and great-power calculations

South Korean military projections assume conscription pools that won’t exist by the 2040s. Russia’s demographic problem has quietly driven its strategic anxiety for two decades. China’s working-age population peaked in 2012 and has been shrinking ever since; the much-discussed demographic crossover with India in 2023 is part of a deeper structural shift in Asian power balances that was set in motion by birth rates in the 1990s. The geopolitics of the 21st century will be shaped by who has and who doesn’t have people.

The meaning of “wealth” begins to change

When real estate stops appreciating because there are fewer buyers in every age group, when stock market returns depend on consumer base growth that won’t happen, and when pension wealth becomes mathematically unattainable, the cultural meaning of being “rich” will detach from the assets that have historically defined it. Japan has been the leading laboratory for this phenomenon since the 1990s, and the Western consensus on what is “weird about Japanese economics” will likely age into “ah okay, that’s just what aging societies look like.”

Growth-based legitimacy is cracking

Since 1945, democracies and authoritarian regimes alike have justified their existence through the promise of “next year being better than this year”. When that promise can’t be kept structurally, the political effects are unpredictable. Some countries adapt by importing labor and then negotiate the resulting cultural anxiety with mixed success. Others double down on growth-restoration mythologies, such as pro-natalist nationalism, ethnic-replacement panic, and neo-traditionalist reaction. The political turbulence of the late 2010s and 2020s is partly about this, even when it pretends to be about other things.

None of this is hypothetical. Most of it is happening now, just unevenly distributed across the world.

The dystopia we expected vs the one we’re getting

A half-century of science fiction has trained us to expect a Blade Runner-like Los Angeles, an Akira-like Neo-Tokyo, a Soylent Green-like 40-million-person 2022 NYC, or a Cyberpunk 2077-like densely stacked Night City – a world in which humanity has overflowed the planet and is stacked into mega-arcologies, fighting over diminishing resources.

However, the math has politely declined this future. Even in the pessimistic SSP3 scenario (12.6 billion people by 2100), Earth’s population density would be lower than Germany’s today. The “overcrowded planet” has always been more about resource consumption and unequal distribution than raw numbers. The Stockholm Resilience Centre’s 2023 update found that we have already crossed 6 of 9 planetary boundaries – climate change, biosphere integrity, land system change, freshwater, biogeochemical flows, and novel entities – with ocean acidification crossing the line in 2025. None of these breaches are caused by population size alone. They are caused by per-capita consumption that is catastrophically unevenly distributed. A US citizen’s lifetime carbon footprint is roughly ≈40-50 times greater than a Niger citizen’s.

So cyberpunk got the numbers wrong, but the vibe is right – the dystopia isn’t density; it’s inequality. The actual 22nd century, if Spears is right, looks closer to a different aesthetic entirely: it’s less like Akira’s Neo-Tokyo. It looks more like Hayao Miyazaki’s Spirited Away bathhouse at 4am: lit, functional, and weirdly empty. Picture those drone shots of Detroit’s abandoned neighborhoods, but cleaner and scaled across whole prefectures and provinces. We’ll see high-functioning AI cleaning the windows of buildings nobody lives in.

And the space colonization dream?

Here, the depopulation narrative intersects with another 20th-century futurist fantasy: humanity becomes multi-planetary. There are O’Neill cylinders in cislunar space, a self-sustaining Mars colony, and eventually, Dyson spheres and an interstellar diaspora. Musk’s stated goal is to establish a self-sustaining city of 1 million people on Mars by ≈2050-2075.

But it costs ≈1 billion USD per ton of useful payload to reach the Martian surface. Musk’s stated requirement for a self-sustaining city is ≈100,000 USD per ton – a 10,000× improvement that may or may not happen. Most non-Musk aerospace estimates predict the first crewed Mars landing in the mid-2030s, an outpost (not a colony – an outpost) in the 2040s-2050s, and a self-sustaining settlement in the 2060s – if at all – possibly never within this century. In February 2026, Musk announced a delay of 5-7 years, refocusing on the Moon first.

But the more interesting collision occurs at the demographic level. The vision of space colonization is based on the unspoken assumption that humanity will continue to produce humans. That there will be young engineers to design the next generation of spacecraft and manufacturing-floor workers to build them. That there will be soldiers, scientists, farmers, and tradespeople for the next planet. This assumption is precisely what is being undone right now.

If Spears is right, and the 22nd-century world has a population of 2-4 billion people in deeply aged societies with old-age dependency ratios that no civilization has ever sustained, then there may not be enough human capital for an interplanetary expansion when we’d otherwise be ready for it. Korea is already losing its STEM workforce due to demographics, and Japan is consolidating its aerospace industry because there aren’t enough young engineers. Even the US is facing severe STEM labor shortages that will worsen as boomer retirements accelerate. The robots-and-AI argument carries some weight here – Marois et al. (2021) showed that productivity per remaining worker can offset some of the loss. However, care work, in-person services, and physical-world engineering are exactly where automation has been slowest.

Put plainly: we may be reaching the resource limits for space colonization not in dollars or rockets, but in people.

This isn’t a neat tragedy – it’s a genuine paradox. The civilization that finally lowers launch costs may also be the civilization that runs out of young people who are willing and able to leave Earth. The dream of trillion-person colonies in space will die in the same century that a trillion-person Earth dies. The root cause is the same, but the surface is different.

Olshansky et al, in Nature Aging 2024, also closes the radical life extension escape hatch. Since 1990, the 10 longest-lived populations have seen decelerating life-expectancy improvements. Even at a hypothetical life expectancy of 120 years, Spears et al’s sensitivity tests show that population trajectories barely change because births dominate the math, not deaths.

What I actually think after writing all this

We’re living through the global population peak. Not in 2100 or some distant future, but in our lifetime, almost certainly before 2080. The world that the people reading this in 2076 will inhabit will look more like the world of 2050 than the world of 2100, in absolute headcount terms. Two-thirds of humanity already lives in countries that will start shrinking before mid-century.

The dystopias we worried about are mostly not the dystopias we’re getting. The “overcrowded Earth” trope has aged poorly. The actual challenges do not fit any 20th-century narrative templates and are ones we’re not really emotionally prepared for:

For Vietnam specifically, the aging-before-getting-rich problem is real. According to estimates from the East Asia Forum, Vietnam needs to maintain a GDP growth rate of 5.5-6.5% for the next two decades to reach high-income status before the demographic dividend closes. That’s narrow.

For Germany, it shrinks unless immigration increases significantly. The population has depended on immigration for 53 consecutive years, and the only real political question is whether the country will admit this.

Looking at the Spears spike chart again, we see a 400-year period when over 2 billion people were alive simultaneously. This period is sandwiched between hundreds of thousands of years when there were far fewer people, and possibly hundreds of years when there will be far fewer people again. Statistically, demographically, and in terms of plain numbers on a graph, we are the densely packed crest of a wave that has been building since the Industrial Revolution and will recede on its current trajectory.

We can only estimate this based on what we’ve discovered so far and within the limits of our current models. As Kurzgesagt puts it, the future is a faraway land. However, the data we have right now does not resemble the future Hollywood promised us. It looks like a brief, brilliant, statistically anomalous spike – and we’re standing on top of it. The question for the next 50 years isn’t how to manage growth. It’s how to gracefully dismantle a civilization that was designed for a curve that no longer applies.

Sources and further reading

Hero image: Generated with google/nano-banana-pro on May 8, 2026, at 2:56 PM.