Investment management: private equity, venture and angel funds: full guide

How each claim is supported:

EstablishedPublished resultSurveyVendor claimIllustration

Tap any marked phrase for its source.

Section 1 of 9

How this industry makes money

In one sentence

A fund earns a yearly fee on the money it manages and a share of the profit when it sells. AI touches the work of finding deals, the profits of the companies it owns, and the price it can sell them for.

A fund manager raises money from investors, buys companies, improves them and sells them. It earns a yearly management fee, usually a percentage of the money committed, and a share of the profit above a set return, called carried interest. Fees pay for the firm. Carried interest is where the partners make most of their money.

Five pools of money decide how a fund does.

Pool	What it is	Who controls it
Deal work	The cost of finding, screening and checking companies to buy	The firm
Company profits	The earnings growth the firm creates while it owns a company	The firm and each company's management
Exit price	The multiple of earnings a buyer will pay at sale	Buyers and public markets
Borrowing	The debt used to buy companies, and its cost	Lenders
Fundraising	Investors' willingness to commit to the next fund	Investors, based on results

Venture funds work the same way, with less borrowing and more weight on a few very large wins.

The exit price is the pool AI moved first. Software has been one of private equity's biggest bets for years. At TPG, software is 18% of the private equity business's assets; at Blackstone it is 6% of all assets, by the firms' own figures. Where a buyer thinks AI could do what a company sells, the multiple it will pay falls, whatever the company's current profits.

Next: Where AI touches the economics ›

Section 2 of 9

Where AI touches the economics

In one sentence

AI lowers the cost of deal work and can raise company profits, but the largest effect in 2026 has been on what buyers will pay for exposed companies. Time saved on deal work is only money saved if the firm does more deals or hires fewer people.

Deal work: faster first reads, same judgment. Firms use AI to screen targets, summarize data rooms and compare contract terms. Fund managers rank this as their clearest return from AI. The saving is in analyst time. It becomes money only if the team covers more deals with the same people, or grows more slowly than it would have. Otherwise it is capacity, not cash.

Company profits: real, but mostly costs so far. Firms are pushing their companies to use AI in customer service, sales, pricing and software development. So far the benefit is mostly lower costs, and nearly 40% of fund managers did not expect a material financial effect in 2026. Expect a year or more before it shows in earnings.

Exit price: the largest and fastest effect. Software stocks fell almost 30% in five months to February 2026. That repriced private companies too, because buyers set their price by public comparisons. Some firms are now moving money toward businesses they see as less exposed.

Borrowing: lenders are asking the same question. Private credit funds that lend heavily to software fell further than others. A company whose lender now doubts its future will pay more to refinance.

Fundraising: investors want proof. Investors will ask how a fund judged AI risk in its last deals, and how it created value with AI in its companies. A fund with a clear answer raises money more easily.

Next: What is already happening ›

Section 3 of 9

What is already happening

In one sentence

Large firms have built AI into deal sourcing and asked their companies for measurable AI targets. The clearest results so far are in finding deals; the clearest shock came from the market repricing software.

Who	What they did	Reported result	What the source does not show
EQT (2022)	Built its own system, Motherbrain, which tracks more than 10 million companies to find deals	Pivotal in nine of more than 60 venture investments; one was later sold for $700 million	How the other investments were found, or what the system cost
Vista (2025)	Required each portfolio company to set AI goals and put a figure on the benefit	Goals set as part of yearly planning	The results in dollars
Dealmakers (Deloitte, 2025)	Surveyed 1,000 corporate and private equity dealmakers	86% use AI in deal work; 88% of private equity firms had spent $1 million or more	Whether it improved returns
Portfolio companies (Bain, 2024)	Surveyed investors managing $3.2 trillion	About 20% of portfolio companies had AI in use with clear results	Which kinds of company, or how large the results
Buyers of fund stakes (Bloomberg, 2026)	Priced stakes in tech-heavy funds after the software selloff	Discounts rose from about 5% to as much as 20%	Non-public deals, reported by people familiar with them

Two patterns stand out. The clearest gains so far are in finding and checking deals. And the clearest financial result of AI in private markets was a loss of value, when buyers repriced companies they thought AI could replace.

Next: What could change next ›

Section 4 of 9

What could change next

In one sentence

Watch three things: whether software repricing spreads to more of the portfolio, whether AI makes deal work cheap enough that it stops being an edge, and whether investors start asking for proof of AI value creation. Each has a signal you can check.

Scenario 1: The repricing spreads beyond software. Buyers and lenders start marking down business services, data and information companies too. Assumes: AI tools keep taking on work those companies sell. Signal to watch: discounts on fund stakes rich in business services, and lenders repricing loans in those sectors. Last checked: September 2026. The discounts reported so far are in tech-heavy portfolios. Bank analysts estimate 25% to 35% of private lending is exposed to AI, with defaults as high as 13% in a worst case.

Scenario 2: Deal work stops being an edge. Every firm uses similar AI tools to read deal papers, so speed and coverage no longer set firms apart. Assumes: the tools keep getting cheaper and more alike. Signal to watch: shorter sale processes, and sellers' advisers supplying data rooms built for AI as standard. Last checked: September 2026. Adoption is already near universal among large dealmakers.

Scenario 3: Investors ask for proof. Investors in funds start asking how much AI actually added to company profits. Assumes: after two years of promises, investors want numbers. Signal to watch: AI questions in investors' due diligence questionnaires, and funds reporting AI's effect on earnings. Last checked: September 2026. Some firms ask their companies for AI goals in dollars, but public reporting of results is rare.

All three reward the same preparation: a written test of AI exposure for every company and every deal, and dollar goals for AI in the companies most likely to gain.

Next: Who gains and who loses ›

Section 5 of 9

Who gains and who loses

In one sentence

Large firms with data teams and patient money gain most; firms that bought exposed software at peak prices, with heavy debt, lose most. In venture, money is concentrating in a few names, which squeezes everyone else.

Gain: large firms with their own data and operating teams. They can build tools, share what works across many companies, and ask each company for measured results.

Gain: buyers with cash when others must sell. Funds and investors who can buy stakes at a discount from others under pressure pick up assets cheaply.

Gain: companies AI makes stronger. Businesses with their own data, a physical service or a regulated position are harder for a general AI tool to copy, and some firms are shifting money toward them.

Lose: funds that bought exposed software at high prices with heavy debt. They face lower exit prices and tougher lenders at the same time.

Lose: smaller venture funds competing for AI deals. In the first quarter of 2026, AI companies took 80% of global venture money, and four companies took 65% of all of it. Smaller funds are priced out of the leading companies and must find value elsewhere.

Mixed: the companies in the portfolio. Most are still testing AI, and only about one in five had clear results in 2024. Their managers face pressure from owners to show gains before the tools are ready.

Next: Where to look before you invest ›

Section 6 of 9

Where to look before you invest

In one sentence

Start with an AI exposure check on every company you own and every deal you consider; it is cheap and protects value. Then fund deal-work tools and a measured program in the companies most likely to gain.

Where	Prerequisite	Cost range for a fund managing $500 million to $5 billion	Most common reason it fails	Pitch you will hear
An AI exposure check on every holding and every new deal	A short, repeatable test, reviewed by someone who knows the customers	Mostly staff time, plus $50,000 to $150,000 of outside help in year one	Done once as a slide, not asked before every price	"An AI risk score" that does not show its method
AI tools for deal work	Clean past deal files, a rule on what goes into outside tools, and a named person who checks summaries	$100,000 to several hundred thousand dollars a year	Stalled by data security worries, the top barrier dealmakers report	"Cut diligence time in half"
A measured AI program in selected companies	One owner per company, a goal in dollars, a check each quarter	Budgeted per company, not per fund	Every company told to "do AI", with nothing measured	"One platform for the whole portfolio"
True statements about the firm's own AI use	A record of what the firm actually uses AI for	Mostly compliance time	Marketing claims more than the firm does, which the regulator now checks	"Show investors you are AI-first"

The cost ranges are illustrative and will be replaced with reviewed figures. The order matters more than the amounts: check exposure first, because it protects value you already own.

One test before any contract. Ask the vendor to run its tool on a closed deal and compare its output with what your team produced.

Next: Up and down the chain ›

Section 7 of 9

Up and down the chain

In one sentence

Your investors, your lenders and the companies you own all feel AI, and each passes it back to you: investors through their questions, lenders through the cost of debt, companies through their results.

Investors in your funds. Pension funds, endowments and family offices are asking how managers judge AI risk and create value with it. Their answers shape the next fundraise.

Lenders. Private credit funds that lend to software companies were repriced sharply in early 2026, and the most exposed did worst. Expect lenders to ask about AI exposure before they refinance, and to price it in.

The companies you own. Every portfolio company sits in an industry of its own. The guides for those industries show how AI reaches each one, and are worth sharing with each company's management.

Advisers. The lawyers, accountants and consultants on every deal are changing how they work and bill.

Guide: Legal ›

A good answer

The questions logged, and one agreed answer.

Push back if you hear

“Nobody has asked yet.”

Next: What changed ›

Section 9 of 9

What changed

In one sentence

This guide was first published September 29, 2026. Changes that affect its conclusions are listed here, newest first.

September 29, 2026: first published.

Q1 2026 (reflected at launch): AI took 80% of global venture money, with four companies taking 65%. Affects section 5.

February 2026 (reflected at launch): software stocks fell almost 30% from October 2025, and the discount on tech-heavy fund stakes rose to as much as 20%. Affects sections 2, 3 and 4, and changes the guide's main conclusion: the fastest financial effect of AI on funds has been on exit prices.

November 2025 (reflected at launch): the SEC named advisers' statements about AI as an exam priority for 2026. Affects section 6.

Next: Decision exercise ›