This depends on how the liquidation preference is structured. Sometimes each new Series is made more senior than the prior series, so the last money in gets paid first - in that case if the last investor put in $100m they would get it all out before any of the other investors got paid. In other cases, the different series of preferred stock have a "pari passu" preference, meaning they all have equal priority - so in that case, you'd calculate each investor's liquidation preference as a percentage of the total liquidation preference owed to all investors, and they'd split the available proceeds according to that percentage.
This is correct in an M&A scenario, but usually not in an IPO. In an IPO, the most common treatment of preferred is to force-convert everyone who holds it to the equivalent amount of common stock, and such a term typically appears in later-round VC term sheets.
Not only do public investors dislike having a junior series of common stock out of the gate (Google-style two-class shares notwithstanding), but the special rights that come with VC-type preferred stock (series votes, class votes, rights to appoint directors, anti-dilution, blocking M&A, etc.) are eliminated once there are public shareholders.
[Speaking as a former VC now public investor who builds and sells VC cap table modeling software.]
