By mid-2025, the share of people worldwide who have never been infected with SARS-CoV-2 is almost certainly small, though no one can give a precise number. The problem is not a lack of data but a stack of measurement challenges: asymptomatic infections that were never noticed, antibodies that fade within months, and official case counts that captured only a fraction of true infections. Early in the pandemic, seroprevalence surveys put global exposure in the low single digits. After the Omicron waves tore through populations in 2022 and 2023, most estimates suggest the vast majority of people on the planet had been infected at least once. Some individuals do appear to have dodged the virus entirely, and emerging research points to specific genetic and immunological reasons why that might be the case.
What the Early Seroprevalence Surveys Found
Before vaccines became widely available, researchers tried to figure out how many people had actually been exposed to SARS-CoV-2 by testing blood samples for antibodies. These seroprevalence studies gave a much fuller picture than official case counts, which depended on people seeking out a test. A large systematic review of studies conducted mostly through late 2020 found that the pooled seroprevalence in the general population was around 3.4%, which at the time translated to roughly 264 million people worldwide who had been exposed or infected.1Clinical Microbiology and Infection. SARS-CoV-2 seroprevalence worldwide: a systematic review and meta-analysis A separate meta-analysis pegged the median general-population seroprevalence at about 4.5%, though rates varied enormously between settings, from under 1% in some groups to nearly 60% among residents of long-term care facilities.2PLOS ONE. Global seroprevalence of SARS-CoV-2 antibodies: A systematic review and meta-analysis
Those numbers seem almost quaint now. They reflected a moment when the original Wuhan strain and early variants like Alpha and Delta were circulating, lockdowns and travel restrictions were still in place, and much of the world had limited testing. Most people on earth, at that stage, had genuinely never been infected. The landscape changed dramatically once Omicron arrived in late 2021, spreading far more efficiently and largely evading prior immunity from both vaccination and earlier infection. By mid-2022, national serosurveys in countries with good data, including the United States, the United Kingdom, and South Africa, were showing infection-induced antibody rates well above 50% of the population, and in some age groups above 80%. The never-infected population shrank rapidly with each successive Omicron subvariant wave.
Why Official Case Counts Were Always a Severe Undercount
If you relied on confirmed case tallies reported by governments, you would dramatically overestimate how many people avoided infection. The World Health Organization’s official cumulative count topped out at roughly 770 million confirmed cases by mid-2024, against a global population of about 8 billion. That seems to leave billions uninfected, but those numbers only captured people who tested positive on a lab-confirmed or rapid antigen test and had their result reported. Enormous numbers of infections were missed, especially in countries with limited testing infrastructure and during later waves when many people tested at home or did not bother testing at all.
A major reason for the undercount is that a large share of infections produce no symptoms. One global meta-analysis covering more than 14,000 confirmed cases estimated that about 44% of people who were infected never developed clinical symptoms.3PubMed Central. Asymptomatic SARS-CoV-2 Infection by Age: A Global Systematic Review and Meta-analysis A separate analysis of over 350 studies put the truly asymptomatic rate at about 35%, distinguishing between people who never got symptoms and those who simply hadn’t developed symptoms yet at the time they were tested.4PubMed Central. Asymptomatic SARS-CoV-2 infection: A systematic review and meta-analysis Another large meta-analysis, covering nearly 30 million individuals tested, found that among people with confirmed infections, roughly 40% were asymptomatic.5JAMA Network Open. Global Percentage of Asymptomatic SARS-CoV-2 Infections Among the Tested Population and Individuals With Confirmed COVID-19 Diagnosis The rates were even higher in certain groups, including pregnant women and travelers caught by screening programs.
People with no symptoms had almost no reason to seek a test in most settings, especially once mandatory surveillance programs wound down. Those infections simply vanished from the data. Add in mild cases that people mistook for a cold or allergies, home rapid tests never reported to a health authority, and entire regions where testing was scarce, and the gap between confirmed cases and true infections becomes enormous. The Institute for Health Metrics and Evaluation and other modeling groups have estimated that the true number of cumulative infections by late 2023 was several times higher than the official figure.
The Antibody Waning Problem
Seroprevalence surveys are more informative than case counts, but they have their own blind spot: antibodies from a natural infection do not stick around forever. The antibody most often used to detect prior infection, rather than vaccination, targets the virus’s nucleocapsid protein. This is useful because most COVID vaccines were designed around the spike protein, so nucleocapsid antibodies serve as a marker that someone actually caught the virus. The trouble is that these antibodies fade.
A longitudinal study of convalescent blood donors found that for a majority of participants, naturally occurring anti-nucleocapsid antibodies dropped below the detection threshold of a widely used test within about four months after infection.6Scientific Reports. Longitudinal analysis of antibody decay in convalescent COVID-19 patients People who had higher initial antibody levels saw those levels decline even faster, about 1.7 times quicker than those who started lower. A large U.S. blood donor cohort study confirmed the challenge of using nucleocapsid antibodies for population-level surveillance, finding significant waning over time that complicated epidemiologic estimates.7PubMed Central. Trajectory and Demographic Correlates of Antibodies to SARS-CoV-2 Nucleocapsid in Recently Infected Blood Donors, United States
This matters for the “how many never got COVID” question because someone who was infected a year or two ago might now test negative on a serosurvey. Their immune system still remembers the virus in other ways, through memory B cells and T cells, but those are harder and more expensive to measure at a population scale. Seroprevalence data therefore almost certainly undercount cumulative infections, particularly older ones. When a study says 70% of a population has nucleocapsid antibodies, the real figure for cumulative infection is higher because some fraction of previously infected people have already lost detectable antibodies. Researchers have developed methods to track reinfections by looking for temporary boosts in nucleocapsid antibody levels in people who are tested repeatedly over time, but that approach works only in longitudinal cohorts like blood donor pools, not for one-time population snapshots.8PubMed Central. Detection of SARS-CoV-2 Reinfections Using Nucleocapsid Antibody Boosting
Who Might Have Genuinely Avoided Infection
Despite all the undercounting, there does appear to be a subset of people who have gone through five-plus years of the pandemic without ever being infected. Some of these people were simply lucky, or lived in circumstances that reduced their exposure: remote locations, small households, jobs that allowed strict isolation, personal decisions to avoid crowded indoor spaces. But researchers have also identified biological factors that seem to make certain individuals harder for the virus to infect or more likely to clear it so quickly that it never registers as a productive infection.
One well-studied factor involves the immune system’s HLA genes, which encode proteins that present pieces of a virus to T cells and trigger an immune response. A large study found that the allele HLA-B*15:01 was significantly overrepresented among people who were infected but remained completely asymptomatic. People carrying one copy of this allele were about 2.4 times more likely to have no symptoms at all, and those carrying two copies were more than eight times more likely to stay asymptomatic.9Nature. A common allele of HLA is associated with asymptomatic SARS-CoV-2 infection The reason appears to be that T cells in these individuals can recognize fragments of SARS-CoV-2 that look very similar to fragments of common cold coronaviruses, giving them a pre-existing head start. About one in five asymptomatic individuals in the study carried this allele, compared with about 9% of people who developed symptoms. This is not the same as never being infected, but someone who clears the virus so efficiently they never feel sick might reasonably believe they were never infected, especially if they were never tested.
Pre-Existing Cross-Reactive Immunity
The HLA finding fits into a broader picture of pre-existing immune defenses. Before SARS-CoV-2 even existed, a substantial fraction of the population already had T cells that could respond to it. Research has shown that cross-reactive memory T cells capable of recognizing SARS-CoV-2 proteins are present in up to half of unexposed, pre-pandemic healthy individuals.10Frontiers in Immunology. High frequencies of alpha common cold coronavirus/SARS-CoV-2 cross-reactive functional CD4+ and CD8+ memory T cells are associated with protection from symptomatic and fatal SARS-CoV-2 infections in unvaccinated COVID-19 patients These T cells were trained by previous encounters with the seasonal coronaviruses that cause common colds. They are not perfectly matched to SARS-CoV-2, but they can provide a degree of protection, potentially helping some people fight off the virus quickly enough to prevent a full infection.
It is worth stressing that cross-reactive T cells do not make someone immune. Most researchers believe these cells shift the odds rather than providing a hard barrier. A person with robust cross-reactive T-cell responses might clear a low-dose exposure before the virus replicates enough to cause detectable infection, while someone without those cells, facing the same dose, develops a full-blown case. Over many exposures, this probabilistic advantage could plausibly keep some people in the “never infected” column, particularly when combined with other favorable factors.
ACE2 Variants and the Front Door
SARS-CoV-2 enters human cells by latching onto a protein on the cell surface called ACE2. Natural genetic variation in the gene that encodes ACE2 can affect how tightly the virus’s spike protein binds, essentially changing how easy it is for the virus to get through the front door. Lab studies have identified specific ACE2 variants that reduce binding. One variant, D355N, restricts the spike-ACE2 interaction enough to limit infection both in cell cultures and in animal models.11PubMed Central. Susceptibilities of Human ACE2 Genetic Variants in Coronavirus Infection
Broader surveys have catalogued a range of ACE2 variants with predicted effects on susceptibility. Some variants, like K26R and T92I, increase the binding affinity of the spike protein to ACE2, potentially making people more susceptible. Others, like K31R and E37K, decrease that affinity, potentially making it harder for the virus to gain entry.12Communications Biology. Human ACE2 receptor polymorphisms and altered susceptibility to SARS-CoV-2 These protective variants are individually rare, and the frequency of specific ACE2 variants differs across populations.13Frontiers in Genetics. SARS-CoV-2 Susceptibility and ACE2 Gene Variations Within Diverse Ethnic Backgrounds No one has shown that any ACE2 variant makes a person completely resistant to infection. The realistic picture is that these variants tweak susceptibility on the margins, like slightly narrowing or widening a doorway rather than bricking it shut.
Blood Type and Susceptibility
Early in the pandemic, several studies noticed that people with blood type O seemed slightly less likely to get infected or to develop severe disease compared with people who had type A. The proposed explanation involves the sugar molecules that define ABO blood groups, which sit on the surface of many cells including those in the respiratory tract. One hypothesis suggests that the particular carbohydrate structures on type-O cells give the virus less to grab onto, while the structures on type-A cells might facilitate viral binding.14Medical Hypotheses. The influence of ABO blood groups on COVID-19 susceptibility and severity: A molecular hypothesis based on carbohydrate-carbohydrate interactions A related proposal noted that people with type O blood carry natural antibodies against both A and B antigens, which could provide an extra layer of innate defense against virus particles that had picked up those antigens from the cells of a previous host.15Immunobiology. Why blood group A individuals are at risk whereas blood group O individuals are protected from SARS-CoV-2 (COVID-19) infection
The blood-type effect, if it exists, is modest. Later and larger studies found that the association weakened once confounding factors were better controlled, and no one seriously claims that blood type O prevents infection. At most, it might shift the probability slightly. Given that type O is the most common blood group globally, even a tiny reduction in susceptibility could, across billions of people, contribute a few million to the never-infected column. But this is speculative at the individual level, and nobody should assume their blood type protected them.
Can You Prove You Were Never Infected?
Here is the uncomfortable reality: proving you never had COVID is essentially impossible in 2025. A negative nucleocapsid antibody test does not rule out past infection because, as noted earlier, those antibodies can fall below detection within a few months. A person who was infected asymptomatically in 2022 might have no antibodies and no memory of being sick. T-cell assays can sometimes pick up evidence of past infection that antibody tests miss, but they are not widely available and were never designed for individual diagnostic use. And if you were vaccinated, your immune system made antibodies to the spike protein regardless of whether you were ever infected, so a positive spike-antibody test tells you nothing about natural infection.
The practical implication is that many people who believe they never had COVID probably did. The high asymptomatic rate, especially in younger and healthier people, means that a significant fraction of infections went entirely unnoticed. Add in the Omicron waves, which were extraordinarily good at reaching even cautious, vaccinated people, and the statistical likelihood of remaining entirely unexposed through multiple years of a globally circulating respiratory virus becomes very low.
What About Hybrid Immunity and Breakthrough Infections?
By now, most of the world’s population has some combination of vaccine-induced and infection-induced immunity, often called hybrid immunity. Evidence from pooled studies indicates that all three types of prior immunity, whether from natural infection alone, vaccination alone, or a combination, provide over 90% effectiveness at preventing severe illness regardless of the viral strain involved.16PubMed Central. Impact of Prior COVID-19 Immunization and/or Prior Infection on Immune Responses and Clinical Outcomes When breakthrough infections happen in vaccinated and previously infected people, they are overwhelmingly mild. This does not tell us how many people were never infected, but it shapes the practical significance of the question. Even if a small percentage of the population has genuinely never encountered the virus, their risk profile in 2025 looks very different from that of a naive individual in 2020, because nearly all of them have been vaccinated.
The never-infected population that researchers worry about most is unvaccinated people who also avoided infection, because they have no immune priming of any kind. This group has shrunk to a very small fraction, but it still exists, particularly among young children born after the main vaccination campaigns and among adults in regions with low vaccine uptake. For these individuals, a first encounter with SARS-CoV-2 in 2025 is still a first encounter, though the circulating variants now cause generally milder disease in the absence of underlying risk factors, partly because the virus has evolved and partly because most severe-disease-prone individuals have already had at least one exposure.
Why Omicron Changed the Calculation
The emergence of the Omicron lineage in late 2021 was arguably the single biggest factor in shrinking the never-infected population. Omicron subvariants like BA.1, BA.2, and later BQ.1.1 and XBB were far more transmissible than earlier strains and far better at evading both vaccine-induced and infection-induced antibodies. Laboratory studies showed that some of these subvariants escaped neutralization by nearly all existing monoclonal antibodies and likely reduced the effectiveness of then-current vaccines at preventing infection, though not at preventing severe disease.17Immunity. Atlas of currently available human neutralizing antibodies against SARS-CoV-2 and escape by Omicron sub-variants BA.1/BA.1.1/BA.2/BA.3
The practical effect was a series of massive infection waves that swept through populations regardless of vaccination status or prior infection history. In the United States, CDC serosurveys showed that by February 2022, about 58% of the adult population and 75% of children had evidence of prior infection. Those numbers continued climbing throughout 2022 and 2023. In many countries, Omicron essentially ended the period during which it was plausible that large fractions of the population had avoided infection. The people who made it through the Delta wave uninfected often did not make it through Omicron unscathed.
Estimating the Truly Never-Infected
Precise global numbers are unavailable, but the best available evidence suggests that somewhere between 5% and 15% of the population in high-exposure countries may have genuinely never been infected as of late 2024. In lower-density or more isolated populations, the figure could be higher. These are rough estimates derived from serosurveys adjusted for antibody waning, and they carry wide uncertainty. Some of those people avoided infection through strict behavioral precautions. Others were exposed but their immune systems dealt with the virus so swiftly, thanks to favorable genetics, robust mucosal barriers, or cross-reactive T cells, that the infection never established itself.
The answer to “how many people never got COVID” is therefore not a single number but a range that depends on geography, age, behavior, and biology. What can be said with reasonable confidence is that the group is far smaller than most people assume, much smaller than official case counts would suggest, and shrinking with each year that the virus continues to circulate. The individuals who remain in that group are not necessarily “immune” in any absolute sense. Most of them were likely protected by a combination of luck, circumstance, and incremental biological advantages that tilted the odds in their favor across many potential exposures.