What is private equity?

At its core, private equity is a professionalized form of ownership. A general partner (GP) raises a closed-end fund, draws capital from limited partners as deals close, buys companies that are not (or will not remain) publicly listed, and tries to leave those companies more valuable than they were at entry. The industry spans control buyouts, growth equity, carve-outs, take-privates, and special situations. What unifies the strategies is illiquidity, active ownership, and a contractual obligation to return capital within a fund life rather than hold forever.

Unlike a public equity manager who can buy and sell shares daily, a PE firm typically seeks meaningful influence — often a controlling stake or a board seat with governance rights — so it can change pricing, cost structure, leadership, product mix, or bolt-on M&A. That control is the economic rationale for the illiquidity premium LPs expect: if you cannot mark and trade daily, you need a path to create value that public markets may not chase with the same intensity.

Private equity is also a partnership business. LPs commit capital; GPs source and underwrite deals, oversee portfolio companies, and report performance. Fees and carried interest align (imperfectly) the GP with upside. When people say “PE owns the company,” they usually mean a fund advised by a GP owns it on behalf of a diversified LP base. Understanding that triangle — LP, GP, portfolio company — is the first step to reading any PE story, including how AI budgets get approved in 2026.

Private equity works as a repeating cycle: raise a fund, acquire companies, execute a value-creation plan during ownership, exit the investment, and return capital plus gains to LPs after management fees and carried interest. Each stage has its own craft. Fundraising sets strategy and dry powder. Origination and diligence determine whether the thesis is real. Ownership is where most of the operational work happens. Exit crystallizes whether the underwriting was right.

1. Fundraising. The GP markets a strategy — mid-market buyouts, healthcare services, software, infrastructure-adjacent assets — and LPs subscribe with capital commitments. Capital is typically drawn down over several years as deals close rather than invested on day one. The legal documents (LPA, side letters) define fees, hurdles, key-person clauses, investment restrictions, and reporting. Vintage year matters because entry multiples and exit windows differ by cycle.

2. Acquisition. Teams source proprietary or auction processes, build models, diligence commercial, financial, legal, tax, IT, and increasingly AI cost and data-rights risk, then negotiate purchase agreements. In leveraged buyouts, a slice of the purchase price is financed with debt. Equity from the fund sits at the bottom of the capital structure. See our leveraged buyout guide for the mechanics.

3. Ownership. Post-close, the board and management install a 100-day plan and a longer value-creation agenda: pricing, procurement, working capital, talent, product, geographic expansion, and add-on acquisitions. In 2026, many ownership agendas also include an AI workstream — inference cost baselines, use-case pilots, and governance — covered in our AI in private equity playbook.

4. Exit. Sponsors sell to a strategic buyer, another PE firm, or list via IPO when markets allow. Proceeds flow through the fund waterfall: return of capital, preferred return to LPs, then catch-up and carry to the GP per the LPA. Distributions, not mark-to-market optimism, are what LPs ultimately care about.

Textbooks still list three levers: earnings growth (EBITDA), multiple expansion, and deleveraging. In practice, modern buyout value creation is more operational than financial engineering alone. Earnings growth can come from volume, mix, pricing power, cost takeout, or bolt-on M&A that expands the platform. Multiple expansion depends on exit markets, quality of earnings, growth durability, and whether the buyer believes the next owner can keep compounding. Deleveraging improves equity value when cash flow pays down debt — but only if the company can service that debt through cycles.

The best firms underwrite a specific operating thesis, not a generic “we will grow EBITDA.” That thesis might be professionalizing a founder-led sales motion, rolling up fragmented clinics, migrating a software company to a more efficient cloud and AI cost stack, or expanding into adjacent products with shared distribution. Diligence exists to pressure-test whether the thesis is achievable with the management team you will actually have, not the management team in the CIM narrative.

AI does not replace those levers; it changes the cost and speed of executing them. Document-heavy diligence compresses. Customer support and back-office workflows can deflect cost. Pricing and forecasting tools can lift margin. Inference spend, if unmanaged, can also erode the very EBITDA the thesis promised. That dual nature — AI as value driver and AI as cost center — is why PE desks now track deployment capital on the Deal Wire the way they once tracked only classic LBO activity.

Limited partners and investment committees speak in two primary languages of return. MOIC (multiple on invested capital), sometimes called equity multiple, answers: how many dollars came back for each dollar put in? A 2.5x MOIC means $2.50 of value for every $1.00 invested, before or after fees depending on whether you are quoting gross or net. IRR answers: what annualized rate of return does that cash-flow timing imply? The same 2.5x earned in three years is a much higher IRR than 2.5x earned in eight years.

Sponsors love high IRRs because carry waterfalls and league-table storytelling often emphasize them. LPs care about both: a flashy IRR on a quick flip may contribute little DPI (distributions to paid-in) if the cheque was small, while a solid MOIC over a longer hold may matter more to a pension that needs absolute dollars. Gross vs net further complicates headlines — management fees, deal expenses, and carry can materially separate what the GP quotes and what the LP receives.

For worked numerical examples and when to prefer each metric, read IRR vs MOIC. For jargon across the capital structure and fund terms, use the private equity glossary.

Private equity vs venture capital. VC typically funds earlier-stage companies with high uncertainty, minority rounds, and a portfolio construction that expects many zeros and a few outliers. PE buyouts usually target mature cash flows, use leverage more aggressively, and underwrite control plans with defined exit paths. Growth equity sits between them: later-stage, often minority, less leverage, still growth-oriented. People casually say “PE” for all private capital; precision matters when you underwrite risk.

Private equity vs public equities. Public markets offer liquidity, continuous pricing, and regulatory disclosure. PE offers illiquidity, negotiated information rights, and active ownership. Public investors can express a view in days; PE investors express a view over years and must create the conditions for exit. Governance differs too: a PE board can replace a CEO faster than a dispersed public shareholder base usually can. The tradeoff is concentration risk and J-curve timing — early years show fees and undeployed capital before distributions arrive.

Private equity vs private credit. Credit lenders are paid in interest and fees with downside protections; equity owners absorb residual risk for residual upside. In 2020s dealmaking, the two often meet in the same capital structure: a buyout fund’s equity sits under a unitranche or syndicated loan provided by private credit. Understanding both sides helps you read refinancing risk and exit feasibility.

For a decade, “AI in PE” often meant a slide in a CIM about a chatbot pilot. In 2026 the conversation is capital markets grade: multi-billion deployment vehicles, portfolio-wide inference programs, and operating partners who ask for token-level unit economics the way they once asked for plant-level COGS. The firms that treat AI as a shared operating system — quality floors, routing, vendor policy, FinOps — will compound advantage across dozens of portfolio companies. The firms that buy logos without instrumentation will fund someone else’s margin.

Practically, AI shows up in four PE workflows. Origination teams use models to scan adjacency and summarize targets. Diligence teams extract from data rooms and flag contradictions for humans. Value-creation teams automate support, finance ops, and coding assistance where evals clear a bar. And fund-level programs negotiate model access and govern spend so each portco is not reinventing procurement. Our full playbook lives at AI in private equity; live vehicles and cheques are on the Deal Wire.

None of this changes the definition of private equity. It changes the operating toolkit of ownership. PE is still about buying well, improving companies, and selling better. AI is becoming part of “improving companies” — and part of the diligence checklist for whether a target’s cost structure is investable.

If you are new to the asset class, read this page, then the glossary, then IRR vs MOIC and the LBO guide. If you work in value creation or portfolio operations, go next to AI in private equity and the due diligence guide. If you need market evidence rather than definitions, start on the Deal Wire and keep this page as the conceptual map.

From the outside, private equity can look like a sequence of press releases: fundraise announced, deal closed, exit completed. From the inside, quality is visible in quieter places. Diligence memos that name what would falsify the thesis. Boards that review leading indicators, not only trailing EBITDA. Capital structures sized for a downturn, not only a base case. Management incentives that reward durable cash conversion. Reporting that LPs can reconcile to bank accounts and audited statements. Those habits are what separate disciplined ownership from financial tourism.

The same standard now applies to AI workstreams. A “good” AI program in a PE portfolio has an inventory of use cases, a measured cost per successful outcome, evaluation coverage for high-risk workflows, clear data rights, and an owner who can turn spend off. A weak program has a vendor logo, a pilot demo, and a hope that the next budget cycle will sort itself out. If you are learning PE in 2026, learn both the classic ownership craft and the new operating craft — they are converging on the same investment committee table.

Finally, remember that PE is a competitive market for assets. Entry prices reflect what other sponsors will pay. That means edge increasingly comes from information advantage, operating speed, and shared platforms — including shared AI procurement and shared diligence tooling — rather than from leverage alone. Our league table and pricing index exist to make those edges observable, not mystical.