XFG shows Covid evolving to evade immunity not worsen
The Variant That Swapped Its Way To The Top
A virus is not a single thing. It is a swarm of slightly different copies, and every so often two of those copies meet inside one body and trade whole paragraphs of genetic text. That is what happened with XFG, the variant now dominating Covid-19 cases worldwide. It is a recombinant — a hybrid stitched together from two subvariants of the Omicron lineage known as JN.1, likely created when both infected the same person at the same time. The World Health Organization began formally monitoring XFG in June 2025, and it now accounts for roughly 60 per cent of sequenced Covid-19 cases globally. [1] The mechanics matter more than the label. Lab tests run by Ian Mellis at Columbia University in New York and his colleagues tested a range of covid-19 viruses, including XFG and another variant nicknamed Nimbus, against antibodies found in human populations. Both variants bound to host cells slightly less well than their predecessors — a small loss. But that loss was more than offset by a greater ability to evade Covid-19 antibodies. The clumsiness costs the virus almost nothing; the evasion buys it everything.
That trade is why XFG spread. It is not that the virus became stronger in some absolute sense. It became better at avoiding the specific antibodies that most people carry from past infections and vaccinations. The result is a wider pool of people who can be infected again, even if their immune systems still remember how to fight.
The Gap Between A Wave And A Warning
Danny Altmann at Imperial College London describes the current moment as “weird, uncharted territory.” [1] His framing is precise: it is not as good as we hoped, where Covid-19 would barely register like a common cold. It is not as bad as some colleagues feared, where the next disaster is already brewing. That middle ground is the story. It is also the hardest kind of story to act on, because nothing forces action.
The numbers so far lean toward calm. In the UK, 1638 covid-19 cases were identified in hospital settings in the first week of October, a rise from 1328 the week before. But that is part of an expected seasonal rise, and lower than at the same time last year.
Aris Katzourakis at the University of Oxford points out that real-time assessment of a virus’s impact can be very difficult, because it takes time for something that sweeps through a population to turn into hospitalisations and for those hospitalisations to be recorded. The signal lags the event. By the time the curve bends upward in a meaningful way, the wave has already passed through. This is not a reason to panic. It is a reason to watch the instruments rather than the noise.
Earlier this year, Leena Azhar at Fakeeh College of Medical Sciences in Jeddah, Saudi Arabia, and her colleagues reviewed what is known about XFG, concluding that it exemplifies “quiet” evolution, more consistent with fine-tuning “than a shift toward greater virulence.” The virus is drifting, not lunging.
The Cost Of Looking Away
The tools that would catch a lunge are being packed up. Mellis points out that the amount of sequencing data has fallen dramatically over the last year. Sequencing is how new variants are spotted before they are everywhere — it is the smoke detector, not the fire hose. Fewer sequences mean a longer blind spot between when a variant emerges and when anyone notices.
Altmann adds a second concern: governments around the world have adopted what he calls a lackadaisical attitude to vaccine boosters. He points out that most children who are starting school have never been vaccinated, and it is unclear what the consequences of that may be.
None of this means a more dangerous strain is inevitable. It means the conditions under which one could emerge have not gone away. The difference between now and 2021 is that the surveillance that would give early warning has thinned, and the booster campaigns that would top up protection have slowed.