There is no definitive answer. More than five years after the pandemic began, the origin of SARS-CoV-2 remains one of the most investigated and debated questions in modern science. No smoking-gun evidence has emerged to prove the virus was engineered in a lab, and no direct animal source has been identified in the wild. Both a natural animal-to-human spillover and an accidental laboratory leak remain plausible, and the world’s top scientific and intelligence bodies remain split.
Where the Major Investigations Stand
In June 2025, the WHO’s Scientific Advisory Group for the Origins of Novel Pathogens published its most comprehensive report to date. The group concluded that “the weight of available evidence suggests zoonotic spillover, either directly from bats or through an intermediate host,” but explicitly stated that “all hypotheses must remain on the table, including zoonotic spillover and lab leak.” The group also acknowledged that much of the information needed to fully evaluate all hypotheses has not been provided, largely because Chinese authorities have limited access to key data.
U.S. intelligence agencies are similarly divided. Four agencies and the National Intelligence Council assess, with low confidence, that the virus most likely jumped from an infected animal to humans naturally. One agency assesses, with moderate confidence, that it most likely resulted from a laboratory-associated incident at the Wuhan Institute of Virology, probably involving experimentation, animal handling, or sampling. Three agencies could not reach a conclusion at all, with individual analysts split between the two explanations or viewing them as equally likely.
The phrase “low confidence” in intelligence terms doesn’t mean unlikely. It means the available evidence is too thin or fragmented for analysts to feel certain about their assessment.
The Genetic Clues That Fuel the Debate
SARS-CoV-2 is closely related to coronaviruses found in bats in Southeast Asia. The closest known relative, a virus called BANAL-52 found in Laotian bats, shares about 96.8% of its genetic code with SARS-CoV-2. Another relative, RaTG13 from a bat cave in China’s Yunnan province, shares 96.1%. These are the nearest matches found in nature, and their binding regions differ from SARS-CoV-2 by only one or two key building blocks at the point where the virus latches onto human cells.
But a roughly 3% genetic gap between these bat viruses and SARS-CoV-2 represents decades of evolutionary distance. No one has found the direct ancestor, the virus that is 99% or more identical, in any animal population. That missing link is a central frustration for researchers on both sides of the debate.
The most intensely scrutinized genetic feature is something called a furin cleavage site, a small insertion in the spike protein that helps the virus enter human cells efficiently. SARS-CoV-2 has a unique four-unit insertion at a critical junction in its spike protein that no other known virus in its family (sarbecoviruses) possesses. Similar features exist in other, more distantly related coronaviruses like MERS, so the structure itself isn’t biologically impossible through natural evolution. But its appearance in SARS-CoV-2 is unusual enough to draw attention from both camps.
Some scientists note that this insertion uses a specific genetic spelling (a CGG-CGG codon pair) that appears with only about 5% frequency in similar viruses, making some researchers argue it looks engineered. Others counter that rare codons do occur naturally and that evolution produces unexpected outcomes all the time. Critically, researchers have pointed out that the techniques used in coronavirus research could produce modifications “practically indistinguishable from a rare coronavirus with a naturally emerging” cleavage site. In other words, if someone did insert it in a lab, there might be no way to tell from the genome alone.
Evidence for a Natural Animal Origin
The strongest evidence for a natural spillover centers on the Huanan Seafood Market in Wuhan, where many of the earliest known COVID-19 cases clustered. Environmental samples collected from the market in January 2020 found SARS-CoV-2 genetic material in multiple locations, with one wildlife stall showing a 30% positive rate. That stall’s contaminated surfaces, including a cart, a feather-removal machine, and a ground sample, also contained DNA from raccoon dogs, civets, bamboo rats, hedgehogs, and other animals previously identified as possible intermediate hosts for coronaviruses.
Researchers also reconstructed near-complete genomes of other animal viruses from these same market samples: a raccoon dog virus, a bamboo rat coronavirus, and a civet virus, confirming that these animals were recently present and shedding pathogens in the same locations where SARS-CoV-2 was detected. This mirrors the pattern seen with the original SARS outbreak in 2003, which was traced to civets sold in similar live-animal markets.
The limitation is that no animal at the market has tested positive for SARS-CoV-2 itself. The environmental samples show the virus and animal DNA in the same place, but that’s correlation, not proof that an animal was infected and passed the virus to a person.
Why a Lab Leak Remains Plausible
The Wuhan Institute of Virology, located roughly 10 miles from the Huanan market, housed one of the world’s largest collections of bat coronaviruses and was actively studying them. Researchers there had collected samples from bat caves across southern China, including the cave where RaTG13 was found, and had published research on engineering chimeric coronaviruses to study their potential to infect human cells. A 2018 grant proposal from the institute’s U.S. collaborators described plans to insert furin cleavage sites into SARS-related coronaviruses, though there is no public evidence that this specific work was carried out.
Lab accidents involving dangerous pathogens are not hypothetical. A CDC presentation noted that in a survey of clinical laboratories, 33% reported at least one lab-acquired infection. There is no mandatory reporting system for such incidents, and the CDC acknowledged there are “disincentives to report laboratory-acquired infections in a litigious and somewhat error-intolerant environment.” The original SARS virus itself escaped from labs in Beijing, Singapore, and Taiwan on separate occasions after the 2003 outbreak was contained.
Chinese authorities have not granted independent investigators access to the Wuhan Institute’s database of virus sequences, which was taken offline in September 2019, or to records of staff illnesses in late 2019. This lack of transparency is the single biggest obstacle to resolving the question.
What “Man-Made” Actually Means
It’s worth distinguishing between two very different claims that often get lumped together. One is that SARS-CoV-2 was deliberately designed as a bioweapon. No credible scientific body or intelligence agency supports this. The virus’s genome doesn’t show hallmarks of intentional weapons design, and the global scientific consensus rejects this scenario.
The other is that a naturally occurring bat virus, or a modified version of one, accidentally leaked from a research lab. This is what most serious proponents of a lab origin are actually arguing. Under this scenario, the virus may have been collected from the wild and studied in the lab, possibly subjected to experiments that enhanced its ability to infect human cells, and then inadvertently released through a safety lapse. This distinction matters because the second scenario doesn’t require a conspiracy or malicious intent, just the kind of human error that has a well-documented history in virology research.
The honest answer, as of mid-2025, is that the evidence is genuinely ambiguous. The market data points toward animals. The lab’s research program and the unusual genetic features point toward a possible accident. Until Chinese authorities share the data that investigators have repeatedly requested, or until someone finds the virus’s direct ancestor in an animal population, this question is likely to remain unresolved.