I saw a headline from “Not Very Private Equity” last week that stopped me cold.

• US private equity firms are sitting on 13,325 unsold companies
• At the current pace of exits, it would take 11 years to clear the inventory
• A third of those companies are four to six years old. Another 27% have been held for seven years or longer

Anyone in PE are probably nodding their heads slowly in agreement. This may account for one of the reasons why there is a rush into AI, hoping by simply have AI in the CIM it will make them worth the enterprise value they are needing. That was the thought nearly a year ago. Does it still hold true?

Inventory Problem

The industry is currently facing a massive inventory problem. They have thousands of companies sitting unsold because the era of easy multiple expansion is dead. This has created a dangerous psychological shift in PE. Instead of focusing on preparing assets for sale, firms are focusing on how to manage them indefinitely. Because they have to.

Let’s be blunt: Private Equity is not in the business of running long-term cash flow businesses. Investors did not sign up to fund a perpetual holding company. They signed up for capital appreciation through successful exits. If you are simply sitting on assets because you cannot find a buyer at your target multiple, you are failing. You are essentially becoming a warehouse manager for mid-market companies.

This accumulation of inventory is killing the ability to raise new funds. The “dry powder” is not being deployed; it is being trapped in these stagnant exits. To try and fix this, firms are leaning heavily into AI as an “efficiency” lever. They think if they can manufacture EBITDA through cost-cutting, they will eventually find a buyer.

The Efficiency Trap

This is the Efficiency Trap. You are trying to solve a pricing problem with a margin play. If you use AI to make your service cheaper to deliver, you are simultaneously making it cheaper for the customer to buy. You might hit your EBITDA targets, but you will do so by destroying the very premium that made the enough company worth buying in the first place.

We are already seeing this happen in the consulting sector. Clients are not stupid; they see the automation happening and they refuse to pay the old rates for work that now takes seconds instead of weeks. McKinsey has already had to shift a significant portion of their global fees toward outcome-based models because clients simply will not tolerate paying for “human hours” when an LLM can do it for pennies.

The mistake many operators are making is thinking that these efficiencies represent a competitive advantage. They do not. If you can modernize your workflow over a weekend using a few well-placed APIs and some prompt engineering, so can your competitor. Efficiency is a baseline requirement now. It is the cost of entry. It is not a moat.

It reminds me of the early days of “Cloud.” Every CIM in 2012 had a slide about how we were “cloud-first.” It meant nothing. It was just a way to say you weren’t running your own servers anymore. Adding “AI” to your deck is doing the exact same thing. It is a buzzword, not a value driver.

Value Over Volume

True value creation does not come from finding new ways to reduce costs. You do not grow a business by shrinking it. You grow a particular kind of business by finding new ways to deliver value that customers actually want to pay for. AI should be used to expand what is possible, not just to prune the headcount. If you use it only to cut, you are just automating your way into irrelevance.

The industry is facing a choice: move the inventory or watch it rot. Making hard decisions even if it means selling at a lower multiple is the only way to free up capital and restore the cycle of deployment and exit. Anything else is just managing a slow, expensive decline.

If you find yourself holding onto assets just because the market is soft, you need to make a hard choice. You might have to take a haircut on the multiple just to move the inventory and free up the capital. Anything else is just managing a slow decline.

Do not try to AI your way into a smaller, cheaper business. Innovate. Add value. Give customers something worth paying for.

AI can help get there, but it is not the answer in of itself.

AI Disclaimer: Gemini Nano Banana Pro was used to generate the photo – Boiler Room (2000)