$1.14T

The mechanics matter, because the same mechanics apply to roughly a dozen other deals on the same shelf.

Thoma Bravo had taken Medallia private in 2021 for $6.4 billion. By April of this year, the equity was worth roughly zero. Blackstone, KKR, Apollo, and Antares — the private credit lenders sitting senior to the equity — now own the company. Thoma Bravo and its co-investors wrote off roughly $5.1 billion. The trigger was not a missed quarter or a sudden customer loss. It was the expiration, at year-end 2025, of a Payment-in-Kind relief arrangement that had let Medallia defer its cash interest by piling it onto principal. Once that mechanism switched off, the cash interest bill arrived in full, against revenue and EBITDA that had not grown the way the 2021 underwriting assumed. The capital structure failed. The company kept operating; the equity went to zero.

Medallia was the first big one, and it will not be the last. Across the 456 enterprise software companies that sit in private-equity hands in our register — combined enterprise value $1.14 trillion — there is somewhere in the neighborhood of $46.9 billion in distressed software loans currently in the private credit market, according to Moody's. The two largest concentrations of that distress sit inside Thoma Bravo's portfolio (currently $183 billion in AUM, 80 active portfolio companies) and Vista Equity Partners' (over $100 billion in AUM). Together, those two firms have deployed more than $120 billion into software since 2019. A meaningful share of that capital was deployed during the 2021 vintage, at peak multiples, with deal structures that financed up to half of the purchase price with debt at rates that only worked when interest rates were near zero.

Those rates are no longer near zero. The expansion math underwriting the debt has not held. And the AI cohort underneath the SaaS layer is now compressing the very expansion math the LBO models needed to clear.

This is the reckoning. It is here, it is in the public record, and it has roughly two more years to run.

Private equity bought enterprise software for fifteen years for the reasons enterprise software made sense as an asset class: predictable recurring revenue, ninety-percent-plus renewal rates, customers too embedded to switch, pricing power that survives most macro environments. Those properties were real, and they justified the leverage that PE structurally needs to clear its target returns.

The book grew in waves, the same way the underlying industry did. The pre-cloud roll-ups (BMC, Infor, McAfee, IFS — all sitting at $11–16 billion of enterprise value today, all founded before 2000) were the first cohort. The SaaS-era take-privates — Anaplan, Avalara, Coupa, Zendesk, Smartsheet, New Relic, Proofpoint, Dayforce, Splunk before its public-market exit — were the second. By the 2021 peak, private equity was acquiring more than 450 software companies a year for over $100 billion in total deal value, according to S&P Global Market Intelligence. Most of those deals were financed with debt that ran as high as half of the total transaction value. Thoma Bravo's $12.7 billion acquisition of Dayforce, completed last year, was nearly half-financed with debt.

The Airframe register shows where this capital sits now:

The pattern is clear. PE's exposure is concentrated in the cohorts the AI wave is most directly repricing: 337 companies, $903B of value, founded between 1985 and 2009, every one of them now competing for budget against the AI-native cohort underneath. Roughly $900 billion of enterprise value is sitting on the wrong side of a productivity ceiling that moved out from under it in the last three years.

The financial press has been calling this "the PE software problem" for two years. It is not a problem yet for every name on the list. It is already a problem for an identifiable subset of them.

One thing worth naming directly: the financial press has been citing a $574 billion exposure figure for two years. That number is roughly half the actual exposure once you reconcile against the full register. The bigger problem is the $1.14 trillion, the $46.9 billion in distressed loans, and the dozen-plus names sitting behind the same capital structures.

These are the named PE-held software companies the trade press has flagged, in 2026, as carrying debt structures that depend on growth or refinancing assumptions the current rate and category environment is not delivering. Numbers below come from the Airframe register; debt and lender details from S&P Global Market Intelligence, SaaStr's analysis, and Octus.

That is roughly $66 billion of equity-plus-debt exposure across nine named deals, and it does not include the dozens of mid-sized deals in the same vintage carrying the same structural issues at smaller scales. SaaStr counts at least twelve more deals with combined debt exposure north of $50 billion at meaningful risk over the next two years.

A leveraged buyout of an enterprise software company works on three assumptions: revenue grows modestly, net retention stays above 110%, and the debt can be refinanced when it comes due. All three are bending at once.

Revenue growth has slowed. Across the named PE-held names above, year-over-year revenue growth has compressed from the 15–25% range at the time of the take-private to mid-single-digits or, in some cases, flat. The categories most exposed — procurement, observability, customer support, customer feedback, document management — are exactly the categories where AI-native competitors are now winning the net-new project at the displacement point. The installed base is renewing. The expansion path is gone.

Net retention has compressed. The 110%+ retention number that LBO models depended on assumed that customers would buy more seats each year. Customers are not buying more seats. In many cases, they are renewing at flat seat counts while quietly evaluating AI replacements for entire workflows. Net retention at scale is moving below 100% in categories where it has not been below 100% in a decade.

The refi window has closed. The 2021 vintage of software LBO debt assumed interest rates would normalize lower and that the IPO window would reopen for the exits. Neither has happened. Refinancing the same debt at 2026 rates roughly doubles the interest cost, which is exactly the math that broke Medallia when its PIK relief expired. The companies most exposed are the ones whose debt comes due in 2026, 2027, and 2028. We are now inside that window.

The first-order effect is what happened at Medallia: the equity goes to zero, the lenders take the company, and the PE firm books the loss against the fund. The second-order effect is the one that matters more for the broader market. Once one of these deals clears at zero equity recovery, every LP in every comparable fund starts asking how much of the rest of the book is held against similar assumptions. Distributions slow. Fundraising slows. The number of software buyout funds with vintage year 2025 is already the lowest since 2017. The denominator of the entire PE software model is contracting in real time.

There is roughly $1 trillion of private credit currently extended to PE-backed companies, with software being one of the largest concentrations within it. Blue Owl, Blackstone Credit, Sixth Street, Golub Capital, Ares Capital, and the rest of the direct lending market hold most of it.

These lenders are senior to the equity, with covenants and seniority protections, and they will recover more than zero on most of these deals. But "more than zero" is a different statement than "made whole." When the equity at Medallia went to zero, the senior loan, which had been trading around par a year earlier, was last marked in the high seventies. That is a real loss to the lender, and it is happening across enough names simultaneously that the publicly traded BDCs with material software exposure are now trading below net asset value, constraining their ability to raise new equity at the same time their existing portfolios are being marked down. Premiums on new private credit issuance have widened by roughly 0.83 percentage points since early 2025. BDC bond issuance in Q1 2026 was down 22% year over year.

The cost of the PE software book is bleeding outward into the broader credit market that was supposed to finance the next round of LBOs. The result is a self-reinforcing tightening: existing PE software deals can't refinance at the rates the original underwriting assumed, the credit funds that would normally provide the refi are themselves capital-constrained, and the IPO window that would let the equity get out is closed because the public software cohort is itself being repriced against AI-native challengers.

The system is not broken. It is unwinding on schedule.

Three things have to clear, and they will not clear simultaneously.

The first is the named at-risk deals. Over the next eighteen to twenty-four months, somewhere between eight and twelve more of the deals on the watchlist above are going to resolve. Some will resolve cleanly: the company executes the AI transition, the debt gets refinanced at higher cost but still works, the equity takes a haircut but survives. Others will resolve the way Medallia did, with the equity going to zero and the senior lenders taking control. The mix matters for the overall PE software return profile, but the binary outcome for individual deals is already largely determined by the debt structure and the category exposure. The forensic analysis a PE firm should be running on its own portfolio in 2026 is not "is this asset compounding through the AI transition" — it is "does this asset's debt structure survive the refi cycle even if it stays flat."

The second is the broader category compression. The 337 PE-held companies founded between 1985 and 2009 are sitting in categories that are, on average, being directly competed against by the AI-native cohort. Most of them are not in the headline-distressed bucket. Most of them will not become Medallia. But the expansion math the LBO models depended on is structurally weaker for the entire vintage, which means the exit multiples available to PE will be lower for the entire vintage. The returns on the 2018–2022 software LBO vintage years are likely to be the worst PE software has produced since the 2007–2009 vintage, and the fundraising data is already pricing that in.

The third is the playbook. A new generation of PE software deals is being structured against an entirely different set of assumptions: lower leverage, more equity, AI-augmented operating playbooks that cut headcount aggressively while preserving the product surface, and outcome-based or consumption-based pricing models that match what the AI-native cohort is doing. The firms that come through this cycle intact will be the ones whose 2024–2026 vintage funds were deployed against the new playbook, not the old one. The firms that doubled down on the 2021 playbook in 2022 and 2023 are the ones whose names will be in the next news cycle.

If you sit on an LP's investment committee, the question that matters in 2026 is not whether to allocate to PE software in the next vintage. It is which of your existing vintage commitments are mark-to-mark exposed to the watchlist above, and which of your GPs are using continuation funds to extend the exit window on deals that should be marked down today. Continuation funds are a defensible structural tool. They are also, in this cycle, a way to defer the recognition of losses that have already happened economically. The disclosure regime around continuation funds is thin. The diligence regime an LP should be running on them is not.

If you sit on a deal team at one of the firms holding the watchlist, the question is which of your portfolio companies are still on the legacy seat-based model, and how aggressively you can route the AI transition through the existing product surface before the next refi cycle hits. Zendesk's $200M of AI ARR is the right kind of move; the question is the rate. Coupa's reported agent-first re-architecture is the right kind of move; the question is the rate. The companies that ship a credible AI-native version of their core workflow before the next debt maturity have a path through. The companies that don't, don't.

If you are an operator inside one of the named deals, the question is the same as it would be inside any company on the AI side of this cycle: which of your renewals is at risk, which of your seats are going to get cut by the AI replacement, and how much of your expansion budget is the AI-native challenger already capturing. The answer to that is in the deployment record. Airframe is where that record lives.