That's a figure with some bounce to it:

    Nuclear reactors connected to the grid in 2021 had a median construction time of 88 months. (7.3 years) [1]

and

    According to an International Atomic Energy Agency (IAEA) study, Tuesday, 15 countries have built a total of 83 nuclear plants over the last 20 years among the 31 countries with nuclear power. It took on average 190 months to build each plant.

    During that period, Korea has built a total of 13 nuclear power plants. The average construction period for each plant was only 56 months (4.6 years), more than three times faster than other countries building nuclear plants. [2] 

South Korea insists it meets Nuclear Safety and Security Commission standards, by contrast:

    In France, [ ... ] taking an average of 126 months to build each plant, nearly three times as long as Japan. [2]

    The United States has a total of 100 nuclear power plants, taking on average 272 months to complete one. [2]

[1] https://www.statista.com/statistics/712841/median-constructi...

[2] https://www.scmp.com/news/asia/article/2027347/south-korea-s...

You should definitely add the timelines for Japan and China from [2]

  Japan, which has built a total of eight nuclear power plants since 1996, was the fastest, taking only 46 months to build each plant, while China ranked third, building 28 nuclear power plants during that period and averaging 68 months to complete each one.

  Japan’s Kashiwazaki-Kariwa Nuclear Power Plant Unit 6 is the world’s fastest-built nuclear power plant, taking only 39 months for completion, while of Korea’s Wolseong Nuclear Power Plant Reactor 3 took 49 months to build.

> The United States has a total of 100 nuclear power plants, taking on average 272 months to complete one.

Looks like one or two plants never finished.

You don't build a nuclear power plant for in ten years though, you build and plan it for the next 100. If there's better sources once it's been built, that's fine, they can be operated alongside the nuclear power plant - or the plant can be decomissioned, it wouldn't be the first one that gets decomissioned even before being put into production.

From the point of view of someone providing a stable grid no matter the cost this is correct. But from an investor's point of view this is problematic.

Most of the cost of nuclear power is in building and demolishing the plant, not in operating it. So if you plan a plant now to go operational in 10 years and run for the next 50, you are locked in to producing electricity for (amortized) $60/MWh for the next decades. Or shut it down early and eat a giant loss, if the projections for solar, wind and storage costs become true. That's a hard sell when you also have the option to build a wind farm in 5 years or a solar installation in two, decommission them after 20 years and repeat with whatever is the most cost effective then.

> if you plan a plant now to go operational in 10 years and run for the next 50, you are locked in to producing electricity for (amortized) $60/MWh for the next decades. Or shut it down early and eat a giant loss

Time will tell if that hypothesis remains true, but inflation is a thing. What you build now with an amortizing cost over decades may weight much cheaper on your debt than you thought.

In that aspect, if you're a state actor, paying more to get lower maintenance costs is usually a good choice.