That's a very confident and incorrect statement.
https://www.nrc.gov/reading-rm/doc-collections/fact-sheets/n...
As for accident vs costs of cleanup, like I said, new plants are designed not only to eliminate accidents from causing failures, but also such that if a failure happened, it would be completely contained.
"The main purpose of the [Price–Anderson Nuclear Industries Indemnity] Act is to partially compensate the nuclear industry against liability claims arising from nuclear incidents [...]."
"By 1966 it had become apparent that the industry would still be unable to obtain adequate private insurance, so the act was extended until 1976. [...] In 2005 it was extended again through 2025 via the Energy Policy Act of 2005."
https://en.wikipedia.org/wiki/Price%E2%80%93Anderson_Nuclear...
Well, not really. While the probability of an accident is indeed low (but not super low by any means, as near misses are often not publicised), the damage/cost of an accident follows a heavy-tailed distribution. As a result the average cost is extremely high.
So high in fact, that there is no market price for it. As a result no nuclear power plant worldwide is insured - the risk is directly born by the society. If we re-distribute the risk to the private sector there would simply be no nuclear power plants.
Regardless if you are pro/contra nuclear - this 'market view' is strong indication that the risk can't be easily dismissed.
Of independent interest: Some studies even conclude that this is one of the rare cases where we observe a distribution without a first moment in real life.