I think when a model is presented like this, it should also show the examples where it fails. In the article, they show Square (Series E valuation at 6 B, pre-IPO 2.66 B) where the "fair valuation" correctly models closer to pre-IPO at 2.2 B.

However, let's look at another example. Take Nutanix (Series E valuation at 2 B, pre-IPO at 2.1 B). This model values it at 0.8 B on their table, almost a third of the IPO price.

There is no explanation forthcoming in this article as to why that's the case. This makes it seem like the Square example was cherry-picked.

I picked NTNX at random, so I don't know if it's the one exception. I'm not going to exhaustively check every result, however. I expect them to do that for me and not sell me a story without pointing out the terrible exceptions.