Long PE holds: where the margin is in years five to seven
Holds now run about seven years. Late in a hold, vendor, agency, software and AI spend is the fastest EBITDA lever, and buyers will diligence it.
In years five to seven of a hold, the fastest margin an operating partner controls usually sits in third-party spend: vendor contracts, agency fees, software licenses and the AI tools that arrived across the portfolio in the last two years. That spend can be renegotiated or cut in a quarter or two, without a new strategy or a reorganization. And a buyer will diligence it anyway, so fixing it before the sale adds to EBITDA and removes a question from the data room.
Why are holds running this long?
Because exits haven't kept up with the backlog. PitchBook counted 33,575 PE-owned companies awaiting exit at June 30, 2026, up from 32,451 at the end of 2025, according to Penn Mutual Asset Management's summary. Bain puts the value of unsold companies at about $3.8 trillion and says holding periods at exit now sit around seven years, against five to six from 2010 to 2021. Almost 40% of portfolio companies have been held for more than five years.
The exit market is moving again. EY counted $212 billion in announced exits in Q3 2026, up 15% from Q2 and 17% from a year earlier. But one good quarter doesn't clear a backlog that size, and LPs are tired of waiting. Bain says distributions were about 14% of NAV in 2025, the fourth straight year below 15%. In EY's survey, 34% of GPs now rate LP pressure for distributions at 8 or higher out of 10, up from 6% two years ago.
So a lot of companies are getting an extra year or two that nobody planned for. The original value creation plan has mostly been executed. The question is what to do with the time.
Why third-party spend, and why now?
Bain's point is blunt: typical deals now need around 10% to 12% average annual EBITDA growth to hit their returns, and that takes real operating work. Late in a hold, most of the big levers are already pulled. Pricing has been reset. The add-ons are integrated. Another reorganization a year before a sale is a risk, not a plan.
Third-party spend is different for three reasons.
- It's controllable. The company signs these contracts. No market has to cooperate.
- It's fast. Renewals, renegotiations and cancellations show up in the P&L within a few quarters, which matters when the next buyer is looking at trailing numbers.
- It drifted. Five years is long enough for a vendor base to grow without anyone looking at it as a whole. Contracts auto-renewed at list price. A second agency was added for a launch and never removed. Two business units bought the same software separately.
None of this is a failure of management. It's what happens when a team is busy executing a plan. But in year six, it's money sitting in plain sight.
Where does AI spend fit?
It's the newest and least governed layer. In EY's Q3 survey, 80% of GPs said AI-related spending across their portfolios rose over the past year, and software models, APIs and vendor licenses were among the largest areas, cited by about half. Only 19% said they're already generating positive net returns from AI. Another 40% said benefits have begun to emerge but haven't been put in financial terms.
That gap is the problem. A buyer will see a growing line of AI subscriptions, API bills and seat licenses with no number attached to the return. In reviews like this, we typically find pilots that became licenses, the same tool bought by several teams on separate contracts, usage-based pricing with no cap, and AI add-ons bundled into renewals of software the company already owned. Each one is small. Across a portfolio, they add up, and they're growing every quarter.
The fix isn't to cut AI. It's to give every tool an owner, a budget and a measure, and to drop the ones that can't show either. We cover the mechanics on our AI spend page.
What does a portfolio vendor desk do?
Most portfolio companies negotiate alone. The same software vendor, cloud provider, telecom carrier or agency holding company sells into several of them, and each one negotiates as if it were the only customer. A portfolio vendor desk changes that. In practice it does four things:
- Maps shared vendors across portfolio companies, with spend, renewal dates and terms side by side.
- Runs a renewal calendar so nothing renews by default, starting each negotiation months before the notice window closes.
- Compares terms across companies. When one company pays more than another for the same product, that's the first conversation.
- Keeps it fixed. New contracts get reviewed before signature, so the savings don't leak back through the next renewal.
The operating partner gets one view of third-party spend across the fund instead of a dozen separate stories. That's the idea behind our Portfolio Vendor Desk, and it sits inside a wider approach to PE value creation.
What will a buyer look for?
In our experience, buyers test the vendor base during diligence, and the questions are getting more specific. Expect questions like these:
- Which contracts renew, reprice or carry change-of-control terms in the next 24 months?
- Are the savings in the adjusted EBITDA already contracted, or still a plan?
- What does the AI and software spend buy, and who owns each line?
- How concentrated are key vendors, and what would it cost to switch?
- Do agency contracts give the company its data, accounts and audit rights?
A seller who can answer these with signed contracts and a clean register defends the number. A seller who can't hands the buyer a reason to discount it. The work also has to happen early. A savings plan presented in the CIM tends to get marked down. Savings that have been in the run rate for two or three quarters get treated as earnings. Our note on the sell-side vendor and cost story goes deeper on what belongs in the data room, and our sell-side readiness work is built around it.
Where should an operating partner start?
Start with the companies most likely to go to market in the next 18 to 24 months. For each, pull 12 months of accounts payable and the contract files, and list every vendor above a set threshold with its owner, renewal date and notice period. Flag anything that renews in the next two quarters. Do the same for AI and software tools specifically, including the ones paid on a credit card.
That gives you the first renewals to work and the first shared vendors to negotiate across the fund. If you want an independent view of where the money is before committing to a full program, start with our 4-week fixed-fee Diagnostic. We're independent: paid only by the client, never by a vendor.
Sources: Penn Mutual Asset Management, Private Equity's Unsold Backlog Keeps Growing as Exit Routes Narrow (citing PitchBook and Bain), Bain & Company, Global Private Equity Report 2026 press release, EY, Private Equity Pulse: key takeaways from Q3 2026
How long are private equity holding periods in 2026?
Bain says buyout holding periods at exit now average around seven years, up from five to six years between 2010 and 2021. Almost 40% of portfolio companies have been held more than five years.
How can a PE portfolio company improve EBITDA late in a hold?
Start with third-party spend: vendor contracts, agency fees, software and AI tools. The company controls these contracts, and renegotiated terms reach the P&L within a few quarters, in time to show in trailing numbers before a sale.
What is a portfolio vendor desk?
A shared function that maps vendors across portfolio companies, tracks renewal dates and compares terms. It negotiates with common vendors as a group and reviews new contracts before they're signed.
Will buyers diligence vendor and AI spend?
Yes. Expect questions on upcoming renewals, change-of-control terms, whether adjusted EBITDA savings are contracted, and what each AI and software line returns. Signed contracts and a clean vendor register defend the number.
We would rather look at your numbers than talk about ours. Tell us the stage of the deal and we will tell you within a day whether the vendor base is worth a look.