What LPs Know About AI That Most GPs Are Still Learning
Coller Capital found that 67% of LPs believe AI will widen return dispersion between GPs, and 70% expect GPs to use it primarily as a cost tool. Those two numbers don't belong in the same world.
- private equity
- limited partners
- AI strategy
- return performance
- Coller Capital
Coller Capital's mid-2026 barometer found two things that don't belong in the same world. Sixty-seven percent of limited partners believe AI adoption will widen return dispersion between leading and lagging GPs: the gap between the best funds and the rest will grow specifically because of how they do or don't use AI. In the same survey, 70% of those same LPs expect GPs to use AI primarily as a cost tool. Only 22% see it as a potential source of return outperformance.
Most LPs believe AI is going to create a meaningful divide between the top funds and the rest. Most of them also expect the funds they invest in to pursue AI primarily for cost efficiency, not competitive advantage. That's not a paradox. It's a description of a market positioning the performance opportunity out of the conversation before it's been properly attempted.
The LP who expects cost efficiency isn't wrong to expect it. Cost efficiency is real, measurable, and reasonably predictable from AI deployment. If a PE fund automates due diligence screening, reduces operating overhead at portfolio companies, or compresses deal execution timelines, there are real savings with real attribution. These show up in fund economics. They're not trivial.
But cost efficiency and return outperformance aren't the same bet. The 5% of organizations generating substantial financial returns from AI, per BCG's September 2025 finding, aren't ahead because they automated more costs. They're ahead because they changed how they create value: redesigned high-volume workflows, built intelligence layers on top of the structured data those workflows produce, and made decisions with information no competitor had access to yet. The output of that work isn't a lower cost structure. It's a structurally different operating model that compounds over time.
That's the prize LPs are under-pricing when they describe AI primarily as a cost tool. And it's the prize most GPs are leaving unclaimed when they structure their AI programs to deliver what LPs expect rather than what's actually available.
There's a two-category problem in PE AI that the Accordion Partners observation from SuperReturn 2026 named directly: there's AI that exists because the board asked about it, and there's AI positioned to show up in the exit multiple. Most portfolios have the first kind. Very few have the second.
The Coller data suggests something about the LP side of that observation: LPs appear to be expecting the first kind. Cost efficiency, operational improvement, expense reduction in the management layer. If that expectation is accurate, the GP who delivers outperformance through AI is doing something most of their peers aren't and most of their investors weren't pricing.
That creates a narrative advantage at exit worth considering. Not because the LP conversation is the primary reason to pursue AI seriously (it isn't) but because the expectation gap is an opportunity. When only 22% of LPs see AI as a potential source of return outperformance, a fund that can demonstrate it is differentiating against a low baseline.
The practical implication isn't about LP communications strategy. It's about what kind of AI program gets built in the first place.
A program designed to deliver cost efficiency asks: which expenses can we reduce? It produces vendor consolidation, headcount normalization, and processing automation. Measurable, defensible, and insufficient.
A program designed to affect the multiple asks: which workflows, redesigned around AI execution, produce an operating margin or revenue advantage that a buyer can independently verify and attribute? It produces structured data assets, intelligence layers, and verifiable EBITDA improvements that compound through the hold period. The question is harder. The answer is what shows up at exit.
Sixty-seven percent of LPs believe AI is going to create a divide. What most of them haven't priced is that the divide is going to be between the funds that asked the first question and the funds that asked the second one.
If you want that question answered for your specific situation, the Forge Playbook does it. Answer a few questions about your business and we'll put together a tailored outline of which workflows are worth automating and what a realistic budget looks like for each. Free, no obligation, takes about three minutes.
