Private equity limited partners spent most of 2026 changing what they trust. PwC's midyear outlook put it plainly: DPI, distributions to paid in capital, has become the more watched metric this year, ahead of IRR, because LPs are no longer willing to accept paper marks as evidence that a fund has actually performed.
Two different disciplines that get described as one
When I built the LBO and DCF model for a UK IT services buyout, I didn't stop at a single valuation method. I triangulated the entry multiple across the LBO's own implied returns and a separately built DCF, specifically to test whether the roughly 8.2x EBITDA entry price was defensible or just a number the sponsor wanted to believe. The two methods converged in a range of about 7.2x to 10.8x, with the actual entry multiple sitting comfortably inside that band.
I'd describe that discipline honestly rather than overstate the connection to what's happening industry-wide right now. Cross-checking a valuation against a second, independently built method at the underwriting stage is a different activity from an LP demanding realized cash distributions instead of paper marks years into a fund's life. Both are expressions of the same underlying instinct, don't trust a number until something independent of the number itself has tested it, but they operate at different points in a deal's life and solve different problems. It would be a stretch to call my triangulation "the DPI mindset," and I'd rather name the shared principle honestly than force a tidier connection than actually exists.
DPI becoming the more watched metric than IRR.PwC, US Deals 2026 midyear outlook
The current market is forcing that verification instinct on the industry from the outside. Cambridge Associates' benchmark data shows a 13.7 percent net median IRR for buyout vintages from 2000 to 2023, with real dispersion, top quartile managers above 19 percent, bottom quartile below 8. An unrealized IRR sitting inside that range tells an LP almost nothing about which side of the distribution they're actually on until distributions confirm it. Bloomberg's private markets desk has also tracked a less comfortable trend this year, some buyout firms using side deals and gagging orders to manage smaller creditors during debt workouts on struggling portfolio companies, exactly the kind of stakeholder conflict my own returns waterfall was built to model transparently across every layer of the capital structure rather than resolve quietly after the fact.
The lesson underneath both examples isn't specific to one market cycle, even if they're not literally the same mechanism. A model's output is only as trustworthy as the second, independent check applied to it, whether that check happens at underwriting through a triangulated valuation or years later through a demand for realized cash. Both principles were true in 2021 when almost nobody wanted to apply them. The market is only now being forced to relearn that, one distribution at a time.