In June, Forbes reported that Thrive Holdings, a company spun out of Joshua Kushner’s venture firm Thrive Capital, is committing $1 billion to buying local accounting practices and rewiring them with AI. It runs the plays through an operating company called Current, formerly Crete Professionals Alliance, which has already pulled together close to 50 local practices, including Larson Gross, and it is reportedly in talks to raise around $2 billion from investors including SoftBank, Altimeter, and D1. What lifts this above the usual roll-up is the name in the cap table. OpenAI took a stake and lent a team to train AI agents on tax work, an arrangement where the model maker’s ownership grows as the accounting portfolio does. By Current’s own account, it ran 7,000 tax returns through its AI last season.

Set the marketing aside and the deal is a clear photograph of the two forces now remaking who owns the accounting profession at the same time. One is private equity money. The other is machine labor. The interesting question is not whether they are coming, because they are already here. It is whether they actually agree with each other.

Start with how far the money has already gone, because it is easy to underestimate. According to research the International Federation of Accountants published this spring, 177 direct private-equity investments in accounting firms between 2015 and 2025 set off another 875 roll-up acquisitions underneath them, more than a thousand impacted firms in total. In 2025 alone, fewer than 200 direct investments spawned roughly 900 follow-on deals, and IFAC’s measure of how much each investment multiplies into further consolidation has risen about fourfold since 2021.

This is no longer happening at the edges of the profession. By early 2026, roughly half of the 30 largest US accounting firms carried some form of private-equity money or an alternative practice structure, the arrangement that lets outside capital in. Citrin Cooperman, one of those top firms, has already been sold from one private-equity owner to another, the sort of sponsor-to-sponsor handoff that signals an asset class maturing. The ownership of the profession is changing hands, quickly, and mostly out of public view. IFAC was blunt about where this is heading, writing that it expects the numbers to climb “much higher once someone perfects the burgeoning AI-forward, roll-up practice model.” Thrive and OpenAI intend to be the ones who do.

The attraction was never mysterious. An accounting firm throws off the kind of revenue investors dream about: recurring, contractual, and stubbornly recession-resistant, because businesses need their books kept and their taxes filed whether the economy is booming or breaking. The industry is also gloriously fragmented, thousands of small partnerships run by owners staring down a succession problem, with too few young accountants coming up behind them and no obvious way to cash out. A buyer with capital can offer those partners an exit, bolt their firms together, centralize the back office, and steer the combined book toward higher-margin advisory work.

There is a legal wrinkle that shapes every one of these deals. In most of the United States, a firm that performs audits has to be owned by licensed CPAs, so private equity cannot simply buy the whole thing. The workaround, the alternative practice structure, splits the firm in two: a CPA-owned entity keeps the attest and audit work, and a separate, investor-owned company takes the tax, advisory, and back-office services and the profits that come with them. It is legal, it is now common, and it means that when people say private equity “bought” a firm, the reality underneath is a little more engineered than the headline.

For years the private-equity thesis was mostly financial. Buy the reliable cash flows, cut duplicated overhead, add scale, sell in five years at a higher multiple. The newer model says a quieter part out loud. Thrive and Current are not only buying cash flows. They are buying firms in order to take the labor out of them with software and keep the difference. OpenAI’s presence on the ownership side is the clearest signal of intent anyone has offered: the company building the models is now also an owner of the firms deploying them, betting the same direction with its own balance sheet.

The bull case is a genuine flywheel, and it deserves to be taken seriously. Scale funds the AI. The AI cuts the cost to serve. Lower cost and faster turnaround win more clients. Those clients fund more acquisitions. And the data flowing through 50 combined firms can train models an independent shop could never build alone. If that loop holds, a well-run AI roll-up becomes very hard for a twelve-person firm to compete with on price. Current’s own reported numbers, a 31 percent time saving on returns and data-entry accuracy it puts far above the typical manual error rate, are exactly the kind of figures that make the flywheel look real. Treat vendor-supplied performance stats with the usual caution, but the direction is not in doubt.

Now the part the pitch decks skip. Most of these deals were underwritten on the recurring cash flows of the old, labor-heavy, billable-hour model, frequently at rich multiples and often with debt on top. AI is dissolving that exact model. We argued a couple of weeks ago that AI collapses the cost to serve and breaks the hour as the unit of value, and the same logic that makes a roll-up efficient also commoditizes the compliance revenue it was assembled to capture. When a service becomes something widely available software can reproduce, the market stops paying a premium for it and starts asking for a discount. That is not a friendly backdrop for an investor who paid a premium.

There is a second problem, and it is the moat. The AI capability Thrive built with OpenAI is impressive, but a close cousin of it is now sold to every independent firm in the country by Ramp, Basis, Digits, and a dozen others we covered in our tools guide. If the efficiency edge is available for a monthly subscription, it is not a moat, it is a rented advantage that your un-consolidated competitor can rent too. And this is not only the worry of an independent newsletter. The Financial Times reports that private-equity executives themselves are now warning that AI threatens the headcount-heavy billable-hour economics behind their law and accounting bets. When the people who made the bet start hedging it out loud, pay attention.

Put the two problems together and the risk sits in plain sight. Private equity bought a fortress of cash flow at a full price, financed it with debt, and is now watching a technology, in some cases funded by the very same investors, lower the fortress walls.

The skeptical read can be pushed too far, though, because AI genuinely does reward scale, and the bulls are not fools. It takes real capital to deploy these tools well, real data to make them accurate, and scarce, expensive talent to run them safely. It takes a balance sheet to absorb a messy transition year that would sink a small firm. If AI raises the minimum size at which a firm can be efficient and safe at once, then consolidation is a rational response to the technology, not merely financial engineering dressed up in an AI costume. The independent owner who treats staying small as a virtue in itself, rather than a strategy, may be choosing the harder decade.

Both things can be true at the same time, and holding them together is the honest position. The roll-up is genuinely risky for the investors who paid peak prices for a model AI is undercutting. And scale is a genuine advantage for whichever firms come out the other side. The next few years will sort out which effect dominates, firm by firm.

For the aging partner being courted, the message is to know precisely what you are selling. It is a multiple on cash flows that AI may compress, frequently wrapped in an earn-out tied to margin targets that depend on an AI rollout you will not fully control. None of that makes the offer bad. It makes it a decision that deserves both stories, the cash-flow story and the AI story, told straight before you sign, because you are selling into a thesis that its own architects have started to question.

For the firm competing against a PE-backed roll-up, the instinct to match their scale is the wrong one. You will lose that race. The winning move is to out-trust them rather than out-buy them. The independent, judgment-forward firm that supervises its AI in the open, rather than using it to quietly thin the work, is exactly what a consolidating and commoditizing market makes scarce. And the efficiency is no longer theirs alone, because the same tools are for rent, so you can match much of their cost structure without carrying their debt.

For the owner staying independent, the two trends point in opposite and useful directions. The minimum scale needed to compete is rising, but the price of the tools that provide that scale is falling fast. You do not need private-equity money to get AI anymore. You need the discipline to deploy it, a human kept accountable for what it produces, and the nerve to price on the value of your judgment rather than the hours it used to take.

Private equity bet that accounting’s cash flows were a fortress worth paying up for, and for a long time that bet looked obviously smart. AI, sometimes funded by the very same investors, is now quietly lowering the walls of the fortress they bought. The next decade of this profession will not be decided by who holds the most capital or ships the most AI. It will be decided by who understood, before the papers were signed, which of the two forces was actually in charge. For a firm owner standing in the middle of it, the choice was never simply to take the money or fight the machine. It is to see clearly what each of them is really worth.

— Footnote

Footnote is an independent publication, with no affiliate links and no vendor paying for placement. It is not professional accounting, tax, legal, or investment advice. Deal and funding details for Thrive Holdings, Current, and OpenAI are drawn from Forbes and other reporting; Current’s performance figures are the company’s own, reported by Forbes, and not independently verified by us. Consolidation figures are from the International Federation of Accountants; the warning that AI threatens private-equity returns is attributed to Financial Times reporting. Our analysis is opinion and is current as of July 2026.