$1.1T: private-equity software financing

Similar financing structures apply to roughly a dozen other deals.

Thoma Bravo took Medallia private in 2021 for $6.4 billion. By April 2026, the equity was worth roughly zero, and senior private credit lenders Blackstone, KKR, Apollo, and Antares owned the company. Thoma Bravo and its co-investors wrote off roughly $5.1 billion. A Payment-in-Kind relief arrangement expired at year-end 2025. It had allowed Medallia to add cash interest to principal; after it expired, the full cash interest bill came due against revenue and EBITDA below the 2021 underwriting assumptions. The capital structure failed while the company continued operating.

Medallia was the first large example; we expect more. The register contains 425 PE-held software companies with combined enterprise value of $1.1 trillion. Moody's reports roughly $46.9 billion in distressed software loans. Distress is concentrated in Thoma Bravo's portfolio ($183 billion in AUM and 80 active companies) and Vista Equity Partners' portfolio (over $100 billion in AUM). The firms have invested more than $120 billion in software since 2019. Much of that capital entered at 2021 peak multiples, with debt financing up to half the purchase price at near-zero rates.

Higher rates and slower expansion have weakened the debt assumptions. AI competitors are also reducing the growth that SaaS buyout models relied on.

We expect those financing pressures to continue for roughly two years.

For fifteen years, private equity favored software's predictable recurring revenue, renewal rates above ninety percent, embedded customers, and pricing power. Those properties supported the leverage used to meet target returns.

Early roll-ups included BMC, Infor, McAfee, and IFS, all founded before 2000 and now valued at $11 billion to $16 billion. SaaS take-privates followed: Anaplan, Avalara, Coupa, Zendesk, Smartsheet, New Relic, Proofpoint, Dayforce, and Splunk before its public-market exit. S&P Global Market Intelligence reports more than 450 annual software acquisitions and over $100 billion in deal value at the 2021 peak. Debt sometimes financed half the transaction. Thoma Bravo's $12.7 billion Dayforce acquisition, completed last year, was nearly half-financed with debt.

The Airframe register shows where this capital sits now:

PE exposure includes 272 companies founded between 1985 and 2009, with $727B of value, competing with AI-native challengers. Roughly $730 billion of enterprise value faces the productivity benchmark established over the past three years.

Financial reporting has followed the "PE software problem" for two years. The financing stress is already evident in a subset of the cohort.

Financial reporting has cited $574 billion in exposure for two years. Reconciliation with the full register puts the PE-held value at $1.1 trillion, roughly twice that estimate, alongside $46.9 billion in distressed loans and more than a dozen deals with similar financing structures.

Trade reporting in 2026 has identified PE-held software companies whose debt structures depend on growth or refinancing assumptions that current rates and category conditions have weakened. Values below come from the Airframe register; debt and lender details come from S&P Global Market Intelligence, SaaStr, and Octus.

The nine named deals represent roughly $66 billion of equity-plus-debt exposure. That excludes dozens of smaller deals from the same vintage with similar structures. SaaStr identifies at least twelve additional deals with more than $50 billion of combined debt at meaningful risk over the next two years.

Software buyouts commonly assume modest revenue growth, net retention above 110%, and refinancing when debt comes due. All three assumptions face pressure.

Revenue growth has slowed. Across the named companies, annual revenue growth has fallen from 15 to 25% at take-private to mid-single digits or flat. AI-native competitors are winning net-new projects in procurement, observability, customer support, customer feedback, and document management. Existing customers renew, but expansion has weakened.

Net retention has compressed. The 110%+ retention assumed in buyout models depended on customers adding seats. Many now renew at flat seat counts while evaluating AI alternatives for whole workflows. Net retention is moving below 100% in categories that stayed above that level for a decade.

The refi window has closed. The 2021 financing assumptions included lower future rates and an open IPO market for exits. Neither has materialized. Refinancing at 2026 rates roughly doubles interest costs, as Medallia experienced when PIK relief expired. Companies with debt due in 2026, 2027, and 2028 face the most immediate pressure.

At Medallia, equity fell to zero, lenders took control, and the PE fund recorded the loss. Comparable write-offs lead LPs to examine similar assumptions elsewhere in their portfolios. Distributions and fundraising slow. The number of software buyout funds from the 2025 vintage is already the lowest since 2017, limiting capital for the model.

Roughly $1 trillion of private credit is extended to PE-backed companies, with software among its largest concentrations. Blue Owl, Blackstone Credit, Sixth Street, Golub Capital, Ares Capital, and other direct lenders hold most of it.

Senior lenders have covenants and priority over equity, but recovery can still fall short of principal. Medallia's senior loan traded around par a year earlier and was last marked in the high seventies. Similar losses have pushed publicly traded BDCs with material software exposure below NAV, limiting new equity raising as portfolio marks fall. New private credit issuance premiums have widened roughly 0.83 percentage points since early 2025. Q1 2026 BDC bond issuance fell 22% year over year.

Existing software deals face higher refinancing costs while the credit funds that would refinance them have less capital. AI-driven repricing in public software also limits IPO exits. These pressures reinforce one another.

We see this as an unwinding of the original financing assumptions.

We expect three developments, at different times.

The first is the named at-risk deals. We expect eight to twelve further watchlist deals to resolve over eighteen to twenty-four months. Some companies may complete an AI transition, refinance at higher cost, and preserve reduced equity value. Others may follow Medallia, losing equity and handing control to lenders. Debt structure and category exposure heavily influence that outcome. Portfolio analysis should test whether the debt remains viable through refinancing if revenue stays flat.

The second is the broader category compression. The 272 PE-held companies founded between 1985 and 2009 face AI-native competitors even when their debt is not distressed. We expect weaker expansion and lower exit multiples across the vintage. Returns from 2018 to 2022 software buyouts could be the weakest since the 2007 to 2009 vintage; fundraising data already reflects that risk.

The third is the playbook. New deals use lower leverage, more equity, AI operating plans that reduce headcount while maintaining products, and outcome or consumption pricing. We expect funds from the 2024 to 2026 vintages using these assumptions to fare better than firms that repeated the 2021 approach in 2022 and 2023.

LP investment committees can examine existing vintage commitments exposed to the watchlist and continuation funds used to extend exits. Continuation funds can be useful, but can also defer recognition of economic losses. Their limited disclosures call for detailed diligence on valuations and underlying assumptions.

Deal teams can identify portfolio companies still using seat-based pricing and assess how quickly their workflows can adopt AI before refinancing. Zendesk's $200M of AI ARR and Coupa's reported agent-first re-architecture illustrate that direction. The timing relative to debt maturity will determine whether the change is enough.

Operators can review renewals at risk, seats that AI might replace, and expansion spending already going to challengers. Deployment records help answer those questions. Airframe maintains those deployment records.