Thoma Bravo's software portfolio. Vista's. Francisco Partners'. Clearlake's. Add them together and you get something like $400 billion in enterprise value assembled between 2018 and 2022 — the era when software companies traded at 15x-25x revenue and every roll-up thesis sounded brilliant over a Zoom pitch. That era is over. The assets are still there. The exit math is not.

Private equity firms don't mark their portfolios to market the way public investors do. They mark them to model — discounted cash flow, comparable company analysis, recent funding rounds. This gives them enormous discretion on valuation. It also means the gap between what a portfolio company is marked at and what it would actually fetch in a sale can be substantial and invisible for years.

The evidence is in the exits that are happening. Anaplan: Thoma Bravo took it private for $10.7 billion in 2022. It is not worth $10.7 billion today — the comps on planning software have compressed by roughly 40% since 2022. Coupa: taken private at $8 billion, same math, same problem. Citrix: $16.5 billion take-private, merged with TIBCO, the combined entity needed a $1 billion capital infusion from lenders in 2024 just to stay current on covenants. These are not isolated. They are the visible tip of an inventory problem.

What PE firms are doing instead of exiting: continuation funds. The pitch to LPs is "we need more time to execute the value creation plan." The reality is "we need more time to wait for multiples to recover, and if they don't, at least the management fees keep flowing." Blackstone's and Vista's continuation fund activity in software has accelerated every year since 2023. This is what you do when you can't sell: you sell to yourself and call it a continuation vehicle.

The LPs are not stupid. They are just captive. The limited partnership agreement gives the GP discretion. The LP can either roll into the continuation fund on the GP's terms or take a liquidity hit by selling to a secondary buyer at a discount. Neither option is attractive. Both are happening at scale.

Here's the uncomfortable math. If software multiples recover to 2019 levels (roughly 8x-12x revenue for premium assets), most of these portfolios are fine. But if they settle at 6x-8x — which is where the public SaaS median has been consolidating — a material portion of the 2018-2022 buyout portfolio is worth less than the debt used to acquire it. Not zero. Not a crisis. But a decade of 6-8% annualized returns that look awfully like fixed income with the liquidity and risk profile of equities.

The smart LPs know this. They're running their own marks against the GP marks and noting the delta. The less smart LPs are finding out at the annual meeting when the distribution pace doesn't match the IRR on the quarterly report. There is a word for this: duration mismatch.

Continuation funds cannot close a valuation gap. They can only move it to a later date with a different denominator. The reckoning is not a crash — it's a slow decade of returns that institutional allocators will compare unfavorably to the S&P, and the comparison will not be close.