In modern financial theory valuation = cost of replicating the position. One can approach that cost from how the company or the employee would replicate it. My point here is that it is much more expensive than just paying the employee cash due to the optionality and the financing costs associated with such.
The wild thing is, if ones accepts modern financial theory, from the eyes of a risk-neutral investor, stock-based compensation is actually much more costly because of imbedded option value and time value of money.
Say one needed to hedge the other side of a 4y employee stock grant. Let's assume the new cliff-less, monthly vest structure that is now market at some of FAANG. The counter-party would need to borrow a large amount of money to buy some fraction of the shares that the employee is likely to vest based on historical data. This also assumes they are just "delta hedging." There is definitely a "negatively convex" situation where if the stock price increases employees are less likely to leave and if it goes down, employees will find a new job that pays market.
Given the immense volatility in earlier stage companies, the counter-party may elect to hedge the gamma exposure as well, meaning they may need to buy calls in the open market against the RSU position.
In my opinion, Netflix has realized this, and given the implications decided to just pay cash. Dollar for dollar, to a well enough capitalized employee, the stock package is much better. This advantage increases with the volatility of the underlying asset. This completely ignores the career risk of working for a failed startup, but the culture in SV seems to minimize that.