And yet, ultimately, Rubin went out and started his own hardware company (Essential Phone) and it didn't work out well.
Maybe Rubin's domain knowledge wasn't really much of a threat after all. In hindsight, one might argue that Android's success hasn't depended on Android quality/innovation nearly as much as it depended on Google's already powerful position in the search market.
Although I'm not taking a position on whether the numbers are correctly reasoned, I think evaluating the value of the noncompete after the fact and as a static value (was his domain knowledge worth $90 M) isn't likely how it was thought about.
It was probably looked at as insurance against a catastrophic tail risk, the existential threat to the whole Android ecosystem. It might have only been a 10% chance that he could have blown up Android by going to a competitor, but insuring against a 10% chance of multi-$B loss was worth $90 M (for example).
Maybe Rubin's domain knowledge wasn't really much of a threat after all. In hindsight, one might argue that Android's success hasn't depended on Android quality/innovation nearly as much as it depended on Google's already powerful position in the search market.