It certainly can vary quite a bit as the graphs in the article show. Seems that cell chemistry, thermal management, and total cycles are the largest drivers of degradation. EV's with generally smaller, more frequently cycled packs, no active thermal management, and cell chemistry not optimized for that use case will not fare well in hot climates. Something with a larger liquid cooled pack, that doesn't experience as many cycles, will fare much better regardless of the chemistry and easily still have useful amounts of capacity after 20 years.
As a data point, this is from an early 2016 EV with 90,000 miles, roughly 50% DCFS, and about half of its life in a temperate climate. Nearing 10 years and just under 3% degradation.
And to think, we haven't even reached the hotter months of July and August in the boreal summer yet.
At least this looking like a neutral period for the ENSO so maybe we won't break as many records this year.