Money isn't value, but what's transferred in game-theoretic interactions isn't value, it's control. Money is control. It's only possible to compare control of different agents, much less so the value they obtain, and it would be particularly hard to compare the value that 10^9 humans obtain with the value one person obtains. Saying that a person obtains 5% of control that the beneficiaries would be willing to give up in return does seem to make sense.

How well could that control be cashed out to improve the world to entrepreneur's preference is a separate unrelated question. Maybe the benefit people receive directly through the innovation would constitute the majority of the improvement in value of the world according to entrepreneur's preference, as compared to other effects of taking the beneficiary action, including the use of obtained money.