'Carried No Interest' warns of a looming 2028-2029 software PE debt crisis accelerated by AI disruption
Mar 17, 2026 with Carried No Interest
Key Points
- Software PE faces a 2028-2029 debt maturity cliff after borrowing heavily against ARR on unprofitable companies when valuations collapse from 9.1x to 3.4x.
- Enterprise customers are abandoning multi-year contracts for one-year deals citing AI uncertainty, erasing the revenue visibility PE sponsors counted on to roll debt.
- Strategic acquirers have stopped buying late-stage software to fund AI CapEx, eliminating the exit path PE firms relied on to avoid default.
Summary
A debt reckoning is coming for software private equity. Carried No Interest, an anonymous markets observer, warns that structural failures in how PE firms financed software acquisitions over the past five years have created a 2028-2029 debt maturity cliff that will be nearly impossible to refinance in an AI-disrupted market.
The core problem is leverage against revenue, not profit. Starting in 2021-2022, PE sponsors borrowed heavily against ARR to juice returns on unprofitable growth software companies. Valuations collapsed. Unprofitable software fell from 9.1x ARR in 2021 to 3.4x ARR in 2022, but the debt structures remained intact. Sponsors bet they could roll or exit before repayment. That calculation no longer holds.
AI acceleration
AI has compressed the timeline for competitive obsolescence and undermined the cash flow stability that made software PE investments defensible.
Enterprise customers are abandoning long-term contracts. The traditional moat of software PE was predictable three-year contracts. Now customers cite AI uncertainty and are signing one-year deals instead. That fundamental shortening of contract visibility makes the debt maturity cliff much sharper and removes visibility to roll or refinance.
Adjacency threats have also expanded. Rippling and similar well-capitalized platforms can now attack market segments they could not three years ago. That is more disruptive to PE portfolio companies than copycat competitors, because it comes from firms with existing customer relationships and engineering talent.
Exit problem
Liquidity has evaporated on all fronts. Strategic acquirers are redirecting capital to AI CapEx and avoiding late-stage software acquisitions clouded by uncertainty. Public markets are discounting software valuations due to the same uncertainty. The appetite for the $15 billion, 15x ARR mega-deals that characterized 2021-2023 is gone.
Continuation vehicles offer a stalling tactic but only defer the problem. Carried No Interest describes the situation as existential for large platform software PE funds. Recovery rates on defaulted debt are likely to be low, especially for unprofitable companies that cannot be milked for dividends.
PE firms and AI
Most PE firms have been slow to integrate AI expertise into deal evaluation and portfolio operations. Partnerships with frontier labs such as OpenAI and Anthropic are not sufficient. PE needs founder-mode leadership that understands AI capabilities at the top of the firm, plus embedded AI talent across deal teams and operations groups. Without that, PE firms cannot accurately price risk or help portfolio companies compete against AI-native entrants subsidized by VC capital.
People are already leaving platform PE funds looking for the exits.
Software special situations
One emerging opportunity is software special situations where creditors take keys from over-levered PE sponsors and restructure portfolio companies under the Constellation model, which prioritizes disciplined pricing and profitability over growth. Firms like Bending Spoons are positioned to exploit this wave. AI makes the restructuring faster and more viable. Equity holders in over-levered deals will likely take zeros.
Carried No Interest calls Meta's acquisition of Manus a steal. The product has solved browser-based LLM context in ways no frontier lab has yet achieved, making it inference-provider-proof. That technical moat justified the acquisition price within Meta's scale.