The 7% number is the long term inflation adjusted average return. But there have been any number of long periods when the average inflation adjusted return was both higher and lower than that. 7% is just a quick rule of thumb.
However, the author is pulling a bit of a fast one. He starts in January 2000, which was near the peak of the dot com bubble. And ends in May 2020, which was near the bottom of the COVID bust. Adjust those dates a little bit and you come up with very different results. Generally it is a good idea to hold some bonds as ballast (as he recommends) but I don't think the results are nearly as stark as he makes them out to be.