A new study suggests the universe and everything in it could end within the Earth's lifespan - less than 3.7 billion years from now - and we won't know it when it happens.
But one expert says the result isn't valid because the researchers chose an arbitrary end point.
The universe began in a Big Bang about 13.7 billion years ago and has been expanding at an ever accelerating rate ever since.
According to standard cosmology models the most likely outcome for the universe is that it will expand forever.
But a team of physicists led by Raphael Bousso from the University of California, Berkeley, claim their calculations show the universe will end.
Writing in the prepublication blog arXiv.org Bousso and colleagues say there's a "measure problem" in the cosmological theory of eternal inflation.
Eternal inflation is a quantum cosmological model where inflationary bubbles can appear out of nothing. Some expand and go on forever, others collapse and disappear again.
These bubbles, each being a universe, pop in and out of existence like bubbles in boiling water.
They argue, in an eternally inflating universe every event that is possible will eventually occur - not just once, but an infinite number of times. This makes predicting when each event will occur impossible, such as the probability that a universe like ours exists.
"If infinitely many observers throughout the universe win the lottery, on what grounds can one still claim that winning the lottery is unlikely?" they write.
Bousso's team have being trying to determine the number of bubbles that exist at any given time and the number of 'observers' in each bubble to come up with the relative frequency of observers that can live in one universe compared to the relative frequency of observers who can live in another universe.
But the "measure problem" makes calculating this value impossible.
According to Bousso and colleagues, the only way to avoid this conundrum is to introduce a cut-off point, which then helps make sense again.
By introducing this cut-off, they say there is "a 50-50 chance of the universe ending in the next 3.7 billion years."
Dr Charles Lineweaver from the Australian National University's Mount Stromlo Observatory says Bousso's team are simply imposing a catastrophe for statistical reasons.
He says the need for a better statistical solution has led the researchers to a false conclusion about the end of the universe.
"Because the problem won't go away in their calculations, they conclude the universe must really end," says Lineweaver.
"Bousso's average life of a universe is a set time, only because that's what happens when you introduce a cut off point to get a reasonable probability."
"It's a statistical technique being taken probably too seriously."