I believe the median endowment of an R1 university is around $1 billion. That corresponds to $30 million to $40 million of sustained income a year. Enough to fund a single reasonably large academic department, or about as much as grant overheads.
Except that universities usually can't use their endowments as they see fit. Big donors often prefer funding fancy buildings and other singular things over daily operations.
Tuition fees are already too low in most public research universities. Combined with the decline of state funding over the decades, many of them are now facing major budget cuts. Take away grant overheads, and you'll start seeing tenured faculty laid off and entire departments closing.
And this is not a perversion of the research funding mechanism but deliberate policy. Back in my previous life, when I was somewhat involved in educational policy in another country, I saw a shift to a similar policy partially inspired by what was already done in the US. The government took a big chunk of core funding for universities, gave the money to research funders, and significantly increased overhead rates. Instead of awarding money based on opaque negotiations between universities and the government, which were prone to corruption, that money would be given according to competitiveness in research.