What Is COVID-19, and Where Did It Come From?

COVID-19 (coronavirus disease 2019) is the disease caused by SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2). The World Health Organization (WHO) named the disease, and the International Committee on Taxonomy of Viruses named the virus, both on 11 February 2020. SARS-CoV-2 is related to the virus behind the 2003 outbreak of severe acute respiratory syndrome (SARS), but it is a different virus, and their genome sequences are 79.6% identical.

The closest known relatives of SARS-CoV-2 are bat viruses. RaTG13, sampled from a bat in Yunnan province, China, in 2013, shares about 96% of its genome. BANAL-52, found in cave bats in northern Laos in 2020, shares 96.8%. WHO’s Scientific Advisory Group for the Origins of Novel Pathogens (SAGO) weighed the evidence on the origin from 2021 to 2025. It judged these bat viruses too distantly related to SARS-CoV-2 to be the direct source of the pandemic.

The virus enters human cells by binding its spike, a surface protein, to a receptor called ACE2. The spike also carries a furin cleavage site, a short stretch that furin, an enzyme plentiful in the lungs, can cut. The site makes it easier for the virus to infect human cells.

No other SARS-like coronavirus is known to have such a site. But similar sites are common in other coronaviruses, including the virus that causes Middle East respiratory syndrome (MERS). The site has been a focus of the debate over the origin.

Infection can pass with no symptoms, cause a mild illness or kill. WHO lists fever, chills and sore throat as the commonest symptoms of the variants now circulating, and most people recover without treatment. Severe COVID-19 can bring respiratory failure, sepsis, blood clots and the failure of several organs. Older adults, people with conditions such as diabetes, obesity or heart or lung disease, and people who are not vaccinated are at higher risk of severe COVID-19.

The virus spreads through the air, in particles that an infected person breathes, talks, coughs or sneezes out. Catching it is likeliest close to that person or in a closed space shared with them. People can pass it on before their symptoms start, or without having any.

The virus has changed as it spread. WHO has tracked five variants of concern: Alpha, Beta, Gamma, Delta and Omicron. Omicron made up over 98% of the virus genomes shared publicly from February 2022 to March 2023, and Omicron strains still make up the vast majority in circulation.

Some people stay ill for months, a condition called long COVID. The US National Academies of Sciences, Engineering, and Medicine define it as a chronic condition present for at least three months after infection. More than 200 symptoms and conditions have been reported with it. How common it is depends on what is counted. WHO puts it at about 6% of people with COVID-19. In a US survey in 2022, 6.9% of all adults said they had ever had it, and 3.4% had it when interviewed.

The earliest known cases clustered in Wuhan, China, in December 2019. On 31 December 2019, WHO was informed of a cluster of pneumonia of unknown cause there. WHO declared a global health emergency, formally a public health emergency of international concern, on 30 January 2020, and called COVID-19 a pandemic on 11 March 2020. It ended the emergency on 5 May 2023, calling COVID-19 an established and ongoing health problem.

By late 2025, more than 7.1 million deaths from COVID-19 had been reported to WHO. The real toll is higher. Deaths were missed because testing was limited, because many countries register deaths poorly, and because overwhelmed health systems caused deaths that were never recorded as COVID-19.

A fuller measure is excess deaths: how many more people died in the pandemic than would have died in a normal period. For 2020 and 2021, WHO estimated 14.83 million excess deaths, about 2.7 times the number reported as caused by COVID-19 in those two years. A separate analysis, by the Institute for Health Metrics and Evaluation (IHME), put the figure higher still. WHO has since estimated the true toll at more than 20 million deaths.

WHO (Msemburi and colleagues, 2023) and the Institute for Health Metrics and Evaluation (IHME; COVID-19 Excess Mortality Collaborators, 2022). IHME counted 5.94 million reported deaths for the same two years. The figure in brackets is IHME’s 95% uncertainty interval.

Some people counted as COVID-19 deaths died with the virus rather than from it, which pushes counts up. But because total deaths ran so far above normal, the larger error worldwide was undercounting, not overcounting.

How deadly the virus was depended overwhelmingly on age. The infection fatality rate is the share of infected people who die. Before vaccines, a systematic analysis estimated how it rose with age. It climbed from a few thousandths of a per cent in children to about one in five among the very old.

COVID-19 Forecasting Team, 2022, for 15 April 2020 to 1 January 2021, before vaccines. A separate analysis (Pezzullo and colleagues, 2023) put the median rate for people under 60 at 0.03%.

That made COVID-19 far deadlier than seasonal influenza, above all for older people. Seasonal flu kills an estimated 290,000 to 650,000 people a year worldwide. By the excess-death estimates, COVID-19 killed many times that number across its first two years.

The risk of dying fell sharply after 2021, as immunity from past infection and vaccination spread and as the Omicron variant took over. COVID-19 vaccines are covered on a separate page.

Two explanations are taken seriously by scientists and investigators. One is a natural spillover: the virus passed from animals to people, probably from bats through an intermediate animal, most likely in the wildlife trade. A second is an accident connected to research, such as a worker infected while studying or collecting a natural virus. A third idea, that the virus was built as a weapon and released on purpose, has no scientific support. The US intelligence agencies agree it was not developed as a biological weapon.

Several lines of evidence point to a natural spillover at the Huanan Seafood Wholesale Market in Wuhan. The earliest known cases of December 2019 clustered around it, including some people with no direct link to the market. Live mammals that can catch SARS-CoV-2, among them raccoon dogs, were sold there in late 2019.

The positive swabs from the market clustered near the stalls that sold those animals. In the positive samples from one wildlife stall, researchers found the DNA of raccoon dogs and other susceptible animals alongside the virus. The earliest viruses fall into two slightly different groups, called lineages, which one analysis reads as the virus jumping from animals to people at least twice. Both earlier coronavirus outbreaks reached people this way: SARS in 2003 through civets, and MERS through camels.

The natural case has real gaps. No infected animal has ever been found, in the market or anywhere else, so the chain from animal to human is not complete. The market may have been where the virus spread widely, not where it first reached a person. The clustering of early cases there may partly reflect where officials looked first. Statisticians have argued that the published analysis of that clustering does not prove the market was the origin.

The case for a research-related accident rests mostly on where the outbreak began and on what is not known. The Wuhan Institute of Virology, a world centre for bat-coronavirus research, is in the city where the pandemic started. US intelligence found that its researchers did some of this work at a modest biosafety level. It reported that several of them fell ill in autumn 2019 with symptoms that could fit COVID-19.

In 2018 a group including the institute had proposed, in a grant application called DEFUSE, to add furin cleavage sites to bat coronaviruses. Some see that as suggestive, because SARS-CoV-2 carries such a site. China has not released the Wuhan laboratories’ records, staff health files or safety audits that WHO and its expert group repeatedly asked for.

The lab case has its own gaps. No laboratory, in Wuhan or anywhere, is known to have held SARS-CoV-2 or a close ancestor before the pandemic, and investigators found no direct evidence of an accident. The reports of ill researchers are not diagnostic, because their symptoms matched ordinary winter illnesses too. Similar furin cleavage sites turn up naturally in other coronaviruses. The DEFUSE proposal was never funded, and the viruses it named are too distant to be the backbone of SARS-CoV-2.

What would settle the question is fairly clear. On the natural side, an infected animal, or a near-identical ancestor virus in the wildlife trade, would close the chain. On the lab side, a record that a Wuhan laboratory held SARS-CoV-2 or a close ancestor before the outbreak would show a research-related origin. Most of what could decide the question is held in China and has not been shared.

WHO’s Scientific Advisory Group for the Origins of Novel Pathogens (SAGO) carried out the main scientific review. In June 2025 it concluded that most of the scientific evidence supports a natural spillover. It also said the origin will stay inconclusive until more data are shared, and that a lab-related accident cannot be ruled out. An earlier joint study by WHO and China, in 2021, had not been able to identify the origin. Within SAGO, three of the original members disagreed with even examining the lab hypothesis, holding that there was no evidence for it.

The US intelligence agencies have never agreed. In 2021 four of them, with the National Intelligence Council, favoured a natural origin, at low confidence. The Federal Bureau of Investigation (FBI) favoured a lab-associated incident, at moderate confidence, and three agencies could not decide. By 2023 the Department of Energy and the FBI leaned to a lab incident, while the Council and four agencies still leaned natural.

On several points the agencies agreed. The virus was not built as a weapon. Almost all judged that it had not been genetically engineered. And none found direct evidence of a lab accident.

The agencies have kept shifting. In January 2025 the Central Intelligence Agency said a research-related origin was more likely than a natural one, at low confidence, CNN reported. According to the same report, the judgement did not rest on new intelligence. In March 2025 the German newspapers Die Zeit and Süddeutsche Zeitung reported that Germany’s foreign intelligence service had judged a lab accident 80 to 95% likely. That 2020 assessment has not been published.

Scientists, asked directly, lean toward a natural origin. A 2024 survey put the question to 168 researchers across 47 countries. On average they gave a natural origin a 77% chance and a research-related accident a 21% chance. Such a survey measures opinion, not evidence.

In Congress, the two parties drew different conclusions. A US congressional subcommittee’s majority concluded in December 2024 that the virus most likely came from a Wuhan laboratory, without presenting new direct evidence. The subcommittee’s Democrats said the origin was still unknown.

Several popular beliefs about COVID-19 are wrong or overstated. That it was no worse than the flu is false: before vaccines it was several times deadlier on average, and far deadlier for older people. That the death counts were inflated is wrong in the opposite direction, because the true toll was undercounted. That the virus was built as a weapon has no credible support, and the US intelligence agencies found it was not.

The idea that the virus came from people eating bats is misleading. Bats are the original reservoir, but the likely natural route runs through the wildlife trade and an intermediate animal, not a meal.

Calling the lab-leak idea a debunked conspiracy theory is also wrong. It was dismissed that way in 2020, but it was never disproved, and the WHO’s own experts cannot rule it out. Yet the US government’s covid.gov address now redirects to a White House page that presents a lab leak as the true origin of the pandemic.

The disease itself, and the harm it did, are settled. COVID-19 is caused by SARS-CoV-2, a bat-derived coronavirus that spreads through the air. It killed more than 7 million people by reported counts. By excess-death estimates the true toll passed 20 million, and the risk fell heavily on older people. It was far deadlier than seasonal flu. It was not built as a weapon, and no sign of genetic engineering has been found.

Where the virus came from is not settled. Most of the scientific evidence, and most surveyed scientists, favour a natural spillover from animals, probably through the wildlife trade near a Wuhan market. That rests on the earliest cases clustering at the market, the susceptible animals sold there, the sign of two separate jumps into people, and the way earlier coronaviruses reached humans.

But no infected animal and no near-identical ancestor virus has been found, and a research-related accident at a Wuhan laboratory cannot be ruled out. The records that could decide the question are held in China and have not been shared. Several intelligence agencies lean toward a lab accident, mostly at low confidence. The honest position, shared by the WHO’s own experts, is that a natural origin is best supported and the question is still open. Anyone claiming certainty either way is going beyond the evidence.

Download the conclusion

The verdict as one image. COVID-19 killed more than 7 million people by reported counts, and more than 20 million by excess-death estimates. Most scientific evidence favours a natural spillover from animals, but the origin is not settled, and a lab-related accident has not been ruled out. The bars are a 2024 survey of 168 scientists. On average they gave a natural origin a 77% chance and a research-related accident a 21% chance. The image is 1,200 by 630 pixels, and anyone may download it and post it.

Download the card (PNG, 1,200 × 630)