The issue is that the later and later stage you get in investing in "startups" the economics change drastically.
IMO the biggest issue with Softbank's or Tiger's model is that they fundamentally did not understand who would buy their investments at a higher price and the amount of money required to do so.
Their entire thesis is built around the idea that public markets are the greater fool here and will pay more than they did. That worked when interest rates were lower and the discounting of earnings was in favor for tech. When that imploded, and tech stocks correspondingly imploded, the greater fool left the game.
Tiger and Softbank were attempting to corner or buy the market to a degree by investing such great amounts that no competitor could compete or win. Unfortunately, they miscalculated the eventual value of winning the market and had already put up such great amounts that there is effectively no way to recover short of another bubble.
Early or earlier stage investing is less susceptible to this math issue because, if you invest in 10 startups and 9 fail, but your fund size is only $100M, it is still possible to recover because the 1 success could legitimately be worth >$100M.
If you are Softbank, and you invest in 10 companies, and maybe 5 fail, but you have put in $200M into each company, at valuations that are in the billions, it gets much harder for the remaining 5 to return the $1B you lost because their valuations are so high already.
Venture is a calculated bet on what the investor after you will pay for a given business AND the loss rate at the stage you are investing at. Both of these variables vary by stage, Tiger and Softbank took this to the Nth degree w/o realizing there was nobody after them.