When a highly pathogenic strain of bird flu strain of avian influenza (H5N1) began sweeping across wild birds and poultry in 2020, it already looked concerning. Five years on, the picture has grown darker and stranger than most would have imagined.
The virus has infected hundreds of millions of farm animals, spilled into mammals at an unprecedented scale, devastated wildlife, and – in the United States – established itself in dairy cattle, a species no one expected to see implicated.
Human cases remain rare. But virologists say the trajectory is troubling, the data patchy and the future uncertain.
“It’s now a global problem,” says Dr Ed Hutchinson, professor of molecular and cellular virology at the University of Glasgow. “As a disease of wild animals, it’s completely out of control. It’s raging around the world, and there’s no feasible containment method other than just watching it infect huge populations of animals.”
How quickly has the virus spread?
The current lineage of H5N1 emerged in Asia in the late 1990s and has since swept from continent to continent. Since 2020, a particularly aggressive branch of the virus – known to scientists as clade 2.3.4.4b – has spread explosively through wild birds.
More than 180 million poultry have been infected in the US alone, and over 1,000 dairy farms have reported outbreaks. Egg prices have soared, and the US government has spent more than $1.19bn (£881m) reimbursing farmers for losses.
For now, the human toll remains limited: only 71 confirmed cases in the US, resulting in two deaths. Historically, though, H5N1 has been far deadlier: since 2003, almost half of all known human infections globally have proved fatal.
Even so, something unprecedented has happened. In early 2024, H5N1 was found in dairy cattle in the United States.
“This was to everyone’s astonishment,” Hutchinson says, “You now have a situation where a large proportion of consumer milk in the US at any given time contains genetic material from these highly pathogenic viruses.”
Thankfully, pasteurisation destroys the virus. But raw milk, along with close contact for workers on dairy farms, still poses significant risks of infection.
That unexpected species jump – along with growing evidence of infection in mammals from seals to foxes to bears – increases opportunities for the virus to change and adapt. Flu viruses are unusually good at evolving because their genome is segmented. If two different strains infect the same host cell, their gene segments can mix and match, producing a new hybrid virus.
This process, known as genome reassortment, has played a key role in sparking past pandemics, and many experts are concerned the same could happen if a person infected with a human influenza were also infected with bird flu. This may, for example, allow the virus to gain the ability to transmit effectively in humans, while maintaining a structure which is alien and undetectable to our immune systems.
“In a sense,” Hutchinson says, “the question about reassortment is not ‘When is it going to happen?’, but ‘Why isn't it happening all the time?’ Because there are so many influenza variants out there, and we don't really know the barriers to it taking off.”
Beyond farms and factories, the outbreak has become a wildlife crisis. In South America, mass die-offs have struck sea lions. And on the Antarctic coast, almost half of the world’s female breeding population of southern elephant seals may already have been killed, according to a recent study.
The US: epicentre and outlier
The US today is one of the most active hotspots for avian influenza transmission. Yet the national response is fragmented. Farm biosecurity measures remain the primary line of defence, even as evidence grows that the virus may be travelling between farms on the wind – a mode that fences, disinfectant and clean boots cannot stop.
Meanwhile, virologists say monitoring and reporting are weakening. US surveillance varies dramatically between states, complicating assessments of how widespread H5N1 infections really are.
Dr Jeremy Rossman, honorary senior lecturer in virology at the University of Kent, says effective containment hinges on broad, coordinated surveillance.
“This requires extensive surveillance of infection in multiple animal populations, as well as monitoring and testing for animal workers at these farms,” he says. “This can allow for management of any infected flocks or herds, rapid identification of human spillover cases and early detection of new mutations in the virus that may affect transmissibility.”
But in the US, Rossman argues, that isn’t happening consistently. “The response is more limited and variable from state to state and farm to farm,” he says. “Without strategic and coordinated surveillance and containment, the risks of a human transmissible H5N1 virus will steadily rise, with unknown, but potentially critical consequences.”
Hutchinson shares this concern. Not only are human and animal infections rising, he says, but data sharing has become patchier and harder to interpret, raising the risk that early warning signs could be missed.
The good news is that effective vaccines already exist to protect both humans and animals – and the US is thought to have millions of doses stockpiled.
How successful is vaccination? France adopted poultry vaccination in 2023 and saw outbreaks fall by 96 per cent in two years, which is impressive. But progress elsewhere has been slow.
In the US, poultry producers and lawmakers have opposed vaccination on trade grounds, warning that it could harm export markets and cost the chicken industry $10bn (£7.4bn) if other countries ban US meat.
This is despite the outbreak already costing the country billions of dollars, not least ordinary people whose grocery bills have risen as egg prices soared 300 per cent in 2 years.
So how worried should we be?
The central fear is human-to-human transmission. So far, it has not happened in any sustained way. But influenza evolves fast, and every new infected host is another throw of the genetic dice.
Rossman warns that high levels of circulation across species increase the odds “of the virus evolving into a strain that has both high transmissibility and high lethality. Given the catastrophic implications if it did, tracking and mitigating the spread of H5N1 virus is of critical importance.”
There are reasons for optimism: vaccines and antivirals should be effective against avian influenza, and the world is far better prepared thanks to lessons learned from COVID, particularly in diagnostics and global surveillance.
But there are caveats too. Scaling vaccines to a global population would take time; antiviral resistance has already appeared in poultry in Canada, and our immunity to H1 and H3 seasonal flu offers little protection against H5 strains.
Perhaps the sharpest warning comes from computer modelling of an outbreak. A 2025 study from Indian researchers found that once a pandemic strain began spreading in humans, the window for effective containment could be just 2 to 10 detected cases. Beyond that, containment would become almost impossible.
What happens next
As 2026 begins, H5N1 is circulating in more species, across more continents, than ever before. It is entrenched in global wildlife. It has destabilised farming. It is mutating, spreading and defying prediction.
Hutchinson’s message is clear: vigilance, not panic. “There are reasons to hope as well as reasons to be alarmed,” he says. “But there’s definitely no reason to be relaxed about this one.”
What scientists want next is simple: more surveillance, more transparency, more vaccination and more urgency, especially in the United States, where the virus is spreading rapidly, and surveillance remains uneven. Whether they’ll get their wishes, however, remains to be seen.
Read more: