AI and Carry: What Happens to Fund Economics When the Work Gets Cheaper

ai in pe carry fund economics governance gp lp alignment value creation Nov 09, 2026

Carried interest is the most elegant fiction in private equity. It converts agents into something that behaves like owners. Without it the asset class collapses into an expensive mutual fund. With it, the industry tells itself a story about alignment that holds together well enough, as long as the fees stay in their lane.

Most private equity professionals are not principals. They are agents working extremely hard to behave like principals, managing other people's money at a captive fund, an independent sponsor, or a large institutional GP. There is nothing dishonourable in that. Keeping them honest about it was always the job of carry.

Which brings us to the question nobody in the industry is asking cleanly. If AI compresses the cost of doing value creation work, who captures the surplus? The answer is not obvious, and three of the four candidates are worse for the GP than the current conversation assumes.

The short answer

Fees keep the lights on. Carry creates the wealth. A firm treating management fees as profit has quietly stopped being a private equity firm and become a fee business wearing a carry costume. AI spending should be judged against that distinction, because it lands in three different layers with three different clocks, and only one of them pays inside the fund's life. The governance rule is one sentence: if an AI initiative cannot be mapped to an increase in future carried interest, directly or indirectly, on a defensible timeline, it should not be approved on the fund's dime. And the honest answer on surplus capture is uncomfortable. Capability that every fund acquires does not become alpha. It becomes a higher entry price.

How high the carry bar actually sits

Start with arithmetic rather than sentiment. A two percent management fee across a six year hold consumes twelve percent of committed capital simply to keep the firm running. For carry to match that fee revenue, the fund has to clear a meaningful bar, and the height of the bar depends entirely on the waterfall.

Exhibit 1 · What the fund must return for carry to equal six years of fees

Waterfall structure	Net profit required	Multiple	Net IRR
Twenty percent carry, no hurdle	60%	1.60x	8.2%
Eight percent preferred return, full catch-up The structure most LPAs actually describe	70.7%	1.71x	9.3%
Eight percent preferred return, no catch-up	118.7%	2.19x	13.9%

Arithmetic under stated assumptions, not a survey. Two percent on committed capital, six year hold, twelve percent cumulative fee drag, eight percent preferred compounded annually at 58.7 percent cumulative. Fee offsets, recycling, deal fees, and transaction expenses are excluded, all of which move the bar in both directions.

Most funds in most vintages do not clear that middle line comfortably. Run the cohort maths and the great majority of carry in the asset class is generated by the top quartile. Which is another way of saying that for most GPs, fees are not keeping the lights on. Fees are the business. The carry costume comes off quietly in the back office while the language in the LPA stays exactly the same.

That is the lens through which AI spending should be evaluated, and almost never is.

Three layers, three clocks

AI in private equity does not operate at one level. It operates at three, and they resolve on completely different timelines. Confusing them is how a firm ends up spending fund money on franchise infrastructure.

Exhibit 2 · Where the money goes, and when it pays

Layer	Typical spend	Clock	Who actually captures it
OpCo	Pricing, procurement, sales effectiveness, working capital, service, forecasting	12 to 24 months	EBITDA moves, the exit moves with it. Drops straight into carry.
Fund	Sourcing, diligence, portfolio monitoring, decision support	Medium, compounding across deals	Real but indirect. Better selection and faster conviction, hard to attribute.
GP	Knowledge management, compliance, fundraising, back office	Long, often past fund life	Franchise value, not fund value. The LP funds it and does not own it.

The alignment cracks in the bottom row. LPs do not invest in the GP franchise. They invest in the fund, and they want AI dollars deployed where the return lands inside their holding period. GP executives own the franchise, and face a standing temptation to spend the same dollars at the layer where benefits accumulate to the firm long after the current fund is liquidated. That is a conflict of interest dressed as a budget decision, and it is invisible unless someone insists on labelling every initiative by layer.

Hence the rule, written as a requirement rather than a principle. If an AI initiative cannot be mapped to an increase in future carried interest, directly or indirectly, within a defensible timeline, it should not be approved on the fund's dime. Not every project needs a six month payback. Some are foundational, some strategic, some create options. But foundation, strategy, and optionality all have to lead somewhere. If the path from an initiative to future carry cannot be drawn on a napkin, the GP is asking the LP to fund infrastructure for a business the LP is not invested in.

Now the harder question: who keeps the surplus

Assume the tooling works. Diligence cycles shorten, monitoring gets cheaper, the analyst layer produces three times the output, and the operating team covers more companies with the same headcount. The cost of the value creation function falls materially. Where does the saving end up?

There are only four places it can go, and the industry conversation has fixated on the one that is least likely.

Exhibit 3 · Four claimants on the same surplus

Claimant	Mechanism	What has to be true for them to win
The GP, through carry	Better selection and faster EBITDA improvement lifts exit value	The capability is genuinely unequal. Advantage lasts only as long as the asymmetry does.
The LP, through terms	Fee pressure once it becomes visible that the cost base has fallen	LPs can observe the cost change. Slow, but this is what allocators are paid to notice.
The seller, through price	More bidders can underwrite complexity, so more assets get competitive processes	Nothing. This happens automatically as the capability diffuses.
Nobody, competed away	Everyone runs the same screens, so entry multiples rise to absorb the improvement	The capability becomes standard. This is the base case, and it is where most of it goes.

The base case is worth sitting with. A capability that every fund can buy from the same handful of vendors is not an edge. It is a cost of participation, like a data room or a quality of earnings report. Both of those improved diligence enormously and neither produced a lasting return premium for anyone, because the improvement was competed into the entry price. There is no reason to expect a different outcome here, and the mechanism is already visible in origination, where the same screens run against the same universe are removing the very thing they were bought to create. That argument is set out in the sourcing machine nobody audits.

The asymmetry that does survive is not the tool. It is whether the firm can act on what the tool produces, which is an operating question rather than a technology one. Two funds running identical screens will diverge entirely on the basis of who has the bench to execute what the screen surfaces, which is the argument for building capability rather than renting it in the internal operator bench.

The three questions to put in the AI budget paper

Everything above compresses into a one page test that any CFO or COO of a fund can apply.

Which layer. OpCo, fund, or GP. Name it explicitly, and state who pays. Fund money at the GP layer requires an explicit conversation with the LPAC, not a line item.

What is the carry path. Draw it. Initiative to operating metric to EBITDA to exit multiple to distribution. If the chain has a missing link, the initiative is a cost, which is a legitimate thing to be, but it should be labelled correctly.

Is the advantage durable or diffusing. If every competitor can buy the same capability within eighteen months, underwrite the benefit as temporary and price it accordingly. The vendor tests that establish whether you are buying a system or a wrapper are set out in diligencing the claim, and the sequencing discipline that stops the budget being written by the wrong function is in the technical veto theory.

The firms that win the next decade will not be the ones with the most tools. They will be the ones that understood, before anyone made them say it out loud, that AI is not a technology initiative. It is a carry augmenter, or it is overhead.

Efficiency is an operational metric. Carry is an ownership metric. If AI is not moving the carry needle, the money is being spent to automate the status of being an agent.

COPPE Level 2

Drawing the carry path, link by link

Initiative to operating metric to EBITDA to exit multiple to distribution. That chain is what COPPE Level 2 builds: nine quantitative modules on one company held from entry to exit, covering the entry model, the debt schedule, the sequencing decisions, and the exit bridge. A supervised certification examination, and three months of PE Simulator access. $349.

Enrol in COPPE Level 2 Or find which programme fits you

VCI Institute in collaboration with Mohamad Chahine
Published 9 November 2026

Related reading from the VCI Institute

The Six Demand-Side Drivers
Where the carry actually gets made, and which of the six a tool can plausibly touch.

The Sourcing Machine Nobody Audits
The clearest live example of a capability being competed into the entry price.

And Then IT Walked In
Why the budget conversation usually starts with the wrong function in the chair.

About the VCI Institute

The VCI Institute is a nonprofit dedicated to building practical capability and shared standards for value creation in private equity. The Institute publishes operator-grade frameworks and runs certification programmes for operating partners, portfolio company executives, and value creation analysts. You can see what each programme actually covers before deciding. Analysis published here draws on the Institute's certification curricula and on structured review of mid-market transaction patterns rather than on any single proprietary dataset. Where a figure is directional rather than measured, it is described as such.

Further material is available in the Institute's Insights library and its free resource library of templates, checklists, and case snapshots.

Affiliated ventures

The Institute is affiliated with two commercial ventures founded by the same team. They are named here rather than inside the analysis above.

Valuethropy builds OPOs, the Operating Partner Operating System: a self-serve system that carries the full value creation role across a mandate, with the human in command at every step.

AIVROS is an AI value creation and readiness platform. It assesses how ready a company, fund, or portfolio actually is for AI, and maps the optimal sequence for adopting AI in service of value creation.

© 2026 VCI Institute. All rights reserved. The frameworks, terminology, and analysis presented in this article are the intellectual property of the VCI Institute. Reproduction or derivative use without written permission is prohibited. Citation with proper attribution is welcomed.