Did humans almost become extinct 900,000 years ago? A biologist explains

According to genetic data published in 2023 study from Scienceour ancestors faced an extreme population bottleneck about 900,000 years ago. This means that just over a thousand breeding individuals remained for over 100,000 years. If true, this would be one of the most severe population crashes ever inferred for a large mammal. In fact, such a severe crash could potentially have erased the human lineage before it really began.

The idea has captured the public imagination because it reframes our evolution. Most would assume, given our success, that it was a steady climb, rather than a narrow escape. However, as with any great scientific claim, it has also sparked heated debate.

Many are now wondering if this was really an event that was destined to disappear – or if what we are actually seeing is a mirage created by the limits of genetic inference. The truth of the matter lies at the intersection of genomics, climate change, and the profound uncertainty of reconstructing life nearly a million years into the past.

Here’s a breakdown of what we know, according to research.

A Bottleneck Hidden in Human DNA

This story began with modern human genomes, not fossils. In the 2023 study, a team of researchers analyzed genetic data from more than 3,000 present-day individuals in both African and non-African populations. Specifically, they used a newly developed statistical method called FitCoal (Fast Infinitesimal Time Coalescent Process). With this, they were able to reconstruct changes in ancestral population size much deeper in time than most previous methods would allow.

The results showed that, between about 930,000 and 813,000 years ago, the actual size of the human population appeared to have dropped to about 1,280 individuals – a decline of more than 98% from previous levels. More surprisingly, the findings suggest that this bottleneck persisted for more than 100,000 years, which is an unusually long time for such a severe demographic collapse.

In evolutionary terms, this means that humans were on the brink of extinction.

An important distinction to note, however, is that the true population size is not same thing as total population or number of people. Instead, it refers to the number of people contributing genes in the next generation: individuals who were able to reproduce successfully. But even allowing for this distinction, this inferred population is still extremely small for a species that has since proliferated across the globe.

Genetics alone do not explain exactly why this congestion occurred. That said, it’s probably no coincidence that the timing coincides with a period of profound environmental upheaval: the Early-Middle Pleistocene Transition.

During this period, about a million years ago, Earth’s climate system was changing dramatically. This shift had a particular influence on the ice cycles, which became longer, colder and much more extreme. The ice sheets expanded and the sea level fell. in turn, ecosystems across Africa and Eurasia were repeatedly disrupted.

For early human ancestors (probably members of the genus Gay earlier Homo heidelbergensis), these changes would be catastrophic. Food sources would have been scarce and their habitats likely fragmented, making survival particularly difficult.

The study authors argue that this prolonged environmental stress may explain why human populations have remained at dangerously low levels for tens of thousands of years. It is also argued that this is why they have not been able to recover as quickly as many species after short-term crashes. And if these findings are correct, then this bottleneck may have shaped the entire trajectory of human evolution.

How people hit the genetic reset button

One of the more interesting implications of the proposed bottleneck is the role it may have played in human speciation. Specifically, the timing of the bottleneck seemingly aligns with when the fossil record becomes conspicuously sparse and ambiguous and is only later followed by the appearance of more recognizable human forms.

Some have speculated that this population crash could have served as a genetic “reset,” in the sense that it may have reduced diversity and set the stage for later evolutionary innovations.

What’s particularly remarkable is that this bottleneck also coincides with estimates of when humans might have lost a pair of ancestral chromosomes. That is, the point in time where we shifted from having 48 chromosomes, like other great apes, to the 46 chromosomes we have today.

While this chromosomal fusion alone did not make us human, it would make it much easier for a small, isolated population to exhibit genetic changes that could be successfully spread and fixed.

An important question many have asked in the wake of the 2023 study is: If we did in fact almost disappear, then why did it take us so long to find out? This is a valid question, the answer to which lies at the limits of traditional demographic models.

Most previous methods have shown little reliability in inferring population sizes beyond a few hundred thousand years. This is because genetic signals from long ago can become blurred by mutation, recombination and subsequent population expansions – especially by the explosive growth of humans over the past 50,000 years. FitCoal was designed to overcome some of these limitations by modeling the genealogical process on much finer time scales.

In simpler terms, rather than averaging over long periods, FitCoal attempts to capture rapid changes in population size, even those buried deep in evolutionary history. This is the methodological advance that enabled the 2023 study to detect a signal that previous analyzes may have missed.

However, new tools also bring new risks.

Did humans really face extinction?

Not all geneticists are convinced that the 900,000-year bottleneck reflects an actual demographic catastrophe. In a next 2024 study published in the journal Geneticsother researchers argued that the signal discovered in the 2023 study could be a statistical artifact: a pattern created by assumptions in the model, rather than a bona fide population crash.

A key concern underpinning this criticism is population structure. Early humans were not a single, well-mixed population, as they probably existed in fragmented groups across Africa, with limited gene flow between them. If a structure like this were indeed missed, FitCoal could wrongly infer a sharp decline in population size.

Another issue is invasion, or gene flow from archaic hominin groups. As further in 2025 research from Molecular Biology and Evolution argues, admixture between different populations can distort estimates of true population size, which would make it appear smaller than it actually was. Critics also point out that the fossil record does not conclusively indicate a near-extinction event at this time, although the fossil record itself is clearly incomplete.

In other words, even if the genetic signal is real, we can’t be 100% sure what it really is within.

So, humanity almost went extinct 900,000 years ago? The most honest answer is: probably, but we don’t know for sure. In 2023 Science The study presents one of the strongest genetic cases ever made for an ancient human strait.

On the one hand, it was methodologically sophisticated, statistically rigorous, and largely consistent with major climate disturbances in Earth’s history. But on the other hand, the claims made push demographic inferences to their limits. Small modeling assumptions can have big effects when reconstructing events that happened nearly a million years ago.

Because this matters to us humans today

If humanity survived an extinction event, then our existence today is the product of an emergency. This would mean that our intelligence, civilization and technology were not as inevitable as we think, but rather that they were mere possibilities that survived a bottleneck that few species escape.

Furthermore, it reframes our resilience as a species. Humans didn’t emerge because we were invincible, but because small populations adapted, endured, and eventually expanded when conditions allowed.

But, more importantly, even if near-extinction isn’t the exact reality at the time—which no one really knows yet—what is nevertheless clear is that early human populations were much more fragile than we once thought. Whether they dwindled to a few thousand individuals or simply endured prolonged hardship, it’s yet another reminder for us to approach human evolution with humility, as it probably wasn’t as smooth a climb as we thought.

How connected is your human mind to the physical world? Take this science-backed test to find out: Nature connection scale

Which animal reflects your most protective human instincts? Take this fun science-inspired quiz to get an instant answer: Animal keeper test