Reading your industry's obituary draft over breakfast.

Hey all,

I dug into the data center water usage question this week after hearing some new angles on a podcast. I felt like I needed to have an opinion, especially living so close to data center alley. After 30 mins, I still didn’t know what to think.Perhaps I shouldn’t be surprised to learn that the citable facts both sides of this debate use depend on entirely different definitions of key metrics.

• Water consumption vs water usage (including recycling)
• Water used per token vs per query vs per chat
• Power costs as peak usage vs average usage\
• Including or excluding power that data centers have contracted to create for themselves as part of the project

The only valuable conclusion I've been able to draw is that, by choosing your interpretation of these metrics, you can pretty much make the story whatever you want it to be.The same thing is happening within portfolio companies that my team works with.

• What do we truly mean by "active customer"?
• How do we define revenue in its different stages and what are those stages called?
• Is our close rate as a percentage of qualified calls or all calls?

You can see how easy it is, even assuming good actors on all sides, for people to get confused about the real truth and for decision paralysis to reign supreme over action and growth.

I want to shout out thanks to the Opus Connect organization, which hosted me for lunch at Zaytinia in DC on Thursday last week. It was a great mix of bankers, PE folks, and service providers. And delicious food.

General consensus about deal flow seems to be that great deals are hard to find right now, and thanks to the difficulty in pricing AI risk, all deals are hard to price with confidence. Absolutely part of the reason for the PE backlog story that you'll see The Guardian picked up on this week.

Enjoy the rest of the memo!

GC

Three Things I Learned This Week

The exit backlog made the front page of a general newspaper

The Guardian ran a piece last week with a headline that would have been unthinkable three years ago. “Private equity faces existential crisis in US as unsold companies pile up.”

The numbers in it are the ones this newsletter has been tracking all year, now written for a general audience. More than 13,500 unsold companies sitting in US PE portfolios, including 2,563 in consumer products and services and 1,536 in healthcare. Hundreds held years past the point funds historically liked to sell. PE-backed companies accounting for the majority of large corporate bankruptcies across 2025 and the first half of 2026.

Jim Baker of the Private Equity Stakeholder Project describing a “record number of unsold companies, many of which they’ve been unable to sell… or at least unable to sell at the prices that they’re looking for.”

The article’s consequence list is Steward Health Care, Saks, Eddie Bauer, Kmart, JoAnn Fabrics. Names people outside our industry recognize, which is the point of it running where it ran.

Reading your industry’s obituary draft over breakfast is uncomfortable. The operational effect is simpler. Buyers who were already cautious now have a mainstream citation for discounting what they cannot verify, and the LPs deciding which managers get the next fund have one more reason to reward cash returns over marks.

A good company is worth what it was worth last month. The proof now has to work harder in more rooms.The way out of the queue is a set of numbers that survive a hostile reader. Read the piece.

Regan Inkster measures the decay two quarters before your board pack does

This week’s podcast guest spent 25 years inside more than 70 Global 1000 companies, three of them as chief architect, before building Phive Dynamics around a strange and useful claim. You can measure organizational coherence from the language a company produces, and it moves roughly two quarters before the financials do.

His dataset is 140-plus companies and ten years of quarterly data, and by his published numbers the coherence score correlated with 52 of 76 financial metrics and ran two to four quarters ahead of them.

What’s important is the asymmetry. Organizations fragment in about 1.9 quarters and recover in 3.8. Decline is twice as fast as repair, which on a five-year hold means a fragmentation you catch late can eat a year of the plan.

The line from the episode that names the whole problem came from a retired telecom executive on Regan’s advisory board. “The problem was really never analytics. There’s a lot of analytics out there. It’s the person at the top, the executive team, the experience of being accountable for a complex system while knowing the information reaching you is incomplete and late.”

His proof cases are not gentle. The FIS and Worldpay merger got a “completely compatible cultures” sign-off from the advisers on both sides. Read through his coherence lens it scored sharply negative before the fact.The outcome was a seventeen billion dollar breakdown inside three years and an exit at roughly a quarter of what was paid.

And in ten years of Starbucks data, one gap, the distance between what the executive team said and what the baristas understood, moved with market cap on a four-to-six-quarter lag. It narrowed every time Schultz came back and widened under every successor.

A company can sit in the exit queue with a green board pack for a year while the thing buyers pay for quietly comes apart, and the standard KPI set reports it two quarters after it happened.

Revenue and EBITDA are reports on the past. In a market where the average hold keeps stretching, a signal that runs two quarters ahead of the P&L is worth taking seriously.

For the deal partner six weeks from signing with a clean quality of earnings and a nagging feeling, his prescription costs nothing. Close the loop between decision and reason.

A short weekly note from the exec team on what was decided and why, cascaded verbatim, and then ask managers two layers down to state the strategy in their own words. What comes back tells you whether the value creation plan is a shared plan or a document.

Watch the full conversation here.

Aging in the queue is not neutral, and the marks prove it

J.P. Morgan Asset Management published the statistic that quantifies what sitting in the queue costs.

Companies held for less than five years have historically exited above the GP’s own valuation estimate 87% of the time. Nearly a third of companies held more than ten years exit below the portfolio mark.

Every year in the queue moves a portco from the first group toward the second. For the long-held, the mark on the books is a ceiling a buyer will test, and the test is diligence.

The queue exists at record scale, sitting in it erodes price, and the erosion shows in the language and the operating seams long before the P&L admits it. The operators who come out of this well are running the diligence on themselves now, while there is still time to fix what it finds.

Two News Stories From This Week in Mid-Market PE and Data

The money is following DPI, and only DPI

Sources: Angel Investors Network on GenNx360 | S&P Global daily update

What happened. GenNx360 closed Fund IV at $865 million, above target, in a market where middle-market fundraising fell more than 40% last year. The stated reason is not subtle. The firm returned $1.3 billion in cash to its LPs in the twelve months before the close.

Across the market, funds above $1 billion took 78.2% of all H1 capital, against 59.1% in 2021, and on average no fund vintage newer than 2016 has yet returned its investors’ paid-in capital.

Why you should care. LPs have stopped paying for paper marks. The managers who can raise are the managers who exit, and that pressure lands directly on portcos as a demand for proof a process would accept.

If your sponsor is mid-raise or pre-raise, expect the questions about which portfolio companies can clear a process to get sharper this quarter. The honest answer requires knowing which numbers hold up under a buyer, which is knowable in advance and mostly is not known.

Europe is consolidating its way toward the exit

Sources: PitchBook Q2 European PE Breakdown via Yahoo Finance | Argos Index via News Asset Pro

What happened. European add-on deals hit 57.7% of all PE deal count in H1, a decade high, and PitchBook’s read is blunt. Sponsors are consolidating because “a consolidated platform is what enhances the chance to exit.”

At the same time the Argos Index put European mid-market valuations at 8.8x EBITDA, a second consecutive quarterly rise off a decade low, and secondaries reached $118 billion in the half, on pace to beat last year’s record.

Why you should care. Every add-on bought to make a platform sellable is an integration bill that comes due at diligence. Customer records that do not reconcile across entities, revenue definitions that differ by acquisition, and a board pack assembled from three charts of accounts on two ERPs are what a buyer’s team finds in week two.

Consolidation as an exit strategy only works if the consolidated numbers agree with each other, and the window to make them agree closes at the banker’s first call.

Free Tool of the Week - The VCP Data Score

The queue above is 13,500 companies long, and every one of them has a value creation plan that assumes the data exists to execute it. Very few have checked. The VCP Data Score is the two-minute version of that check.

It scores a portfolio company's data readiness across the five dimensions a plan depends on, from revenue data through to governance, and hands you one page for the next operating partner review. If a company you back is more than three years into its hold, run it on that one first. The number the score gives you is the number a buyer will find later.

Take the assessment here: 

Sign-off

If any of this lands and there is something you think we can help with, just reply. We read everything that comes through.

As always, forward this on to your favorite PE-backed friend.

Cheers,

Graeme