Private Credit Risk: Why Software Lending Is Under the Microscope

Software companies didn't become private credit's favorite borrowers by accident. They offered subscription revenue, high margins, and customers who rarely left. Artificial intelligence is now putting a question mark next to all three.

That question is why private credit risk has become a sector-specific conversation in 2026. When one industry represents a large share of lenders' portfolios, a change in that industry's economics becomes a portfolio question. Regulators, fund managers, and investors are looking at the same exposure, and so far they are reading it differently.

Why Software Became Private Credit's Favorite Borrower

Lending to software-as-a-service firms grew from about $8 billion in 2015 to more than $500 billion by the end of 2025, roughly 19% of U.S. direct loans, according to Bank for International Settlements (BIS) analysis reported by the ABF Journal.² The Congressional Research Service puts private credit funds' software exposure at around $500 billion as of December 2025.¹

The appeal was structural. Enterprise software tends to come with recurring subscription revenue and high switching costs, which keep renewal rates high.³ ⁴ That predictability changed how some loans were underwritten. Instead of lending against earnings, some lenders sized loans against annual recurring revenue, meaning contracted subscription income, which gave growing borrowers more flexibility. Many of those loans were structured to move to earnings-based covenants over time.³ ⁴

Figure 1: "Private credit loans to software firms."

---

What AI Changes for the Borrower

The assumptions behind those loans were about growth: more customers, more seats, more subscriptions. The concern now is that AI tools could change what customers are willing to pay for. PitchBook reports that early in 2026, investor fears that new AI tools could replace costly subscriptions shifted sentiment toward software lending.⁴ The Congressional Research Service notes that AI coding capabilities have contributed to reduced revenue at some software companies.¹ Bloomberg describes the credit link: many software loans assumed subscriber growth, so even a slowdown in revenue could put pressure on the debt.⁵

The exposure is not uniform. Fitch found that about half of the software companies it rates are at low risk of AI disruption, while 9% are at high risk.⁵ "Software" is a category. Disruption risk is a company-level question.

Private Credit Risk and the Software Concentration Question

The clearest numbers on concentration come from BIS. In a July 2026 bulletin, its economists found that business development companies (BDCs) had lent around $115 billion to software firms, about a fifth of all their lending and over 80% of their technology portfolios.⁶

The same bulletin contains findings that pull in different directions. Revenue uncertainty from generative AI had not yet affected these loans, and neither BDCs nor their equity investors had priced software exposure differently. Credit spreads, meanwhile, had narrowed, which reduces the buffer available to absorb losses. The authors also noted that a few large BDCs are exposed to a shared pool of borrowers, while low leverage and secured lending may limit spillovers.⁶ In effect, lenders were earning less compensation for a risk that had increased.

Credit metrics have moved, but not evenly. Wellington Management reports that median non-accrual rates among publicly traded BDCs rose from 0% in 2022 to 0.33% in the first quarter of 2026, with stress concentrated in certain portfolios rather than spread across the market.⁷

Figure 2: Private credit loans to software firms

Where the Pressure Is Showing Up

Redemptions. The European Central Bank has linked concerns about credit quality and software-sector exposures to a wave of redemption requests from semi-liquid private credit vehicles in the United States. It also notes that non-listed BDCs offer quarterly redemptions at net asset value but are typically subject to gates that can limit or suspend withdrawals during stress.⁸

Banks. Banks lend to private credit funds, which is why supervisors are watching this exposure. In April, the Federal Reserve was reported to be asking major banks for details on their private credit exposure, including the debt funds had taken on from banks.⁹ In early October, Semafor reported that the New York Fed has been visiting several large banks to review those loans, a review prompted in part by one bank's March markdown of loans to software companies threatened by AI. Semafor also notes that such reviews are not uncommon when headlines raise questions about risk.¹⁰

The Treasury's Office of Financial Research struck a similar balance in March. It concluded that vulnerabilities within private credit appear contained, while identifying counterparty exposures between banks and private credit funds as the main channel for risk transmission, one that merits close monitoring.¹¹

[Suggested Visual: Timeline, "2026: how the software question reached regulators." Points: Dec 2025 (about $500B exposure, CRS), March (OFR brief), March (bank markdown of software loans), April (Fed requests exposure details), July (BIS Bulletin 128), August (Fed pilot survey announced), early October (NY Fed bank review reported). Sources: footnotes 1, 6, 9, 10, 11, 12.]

What Remains Unknown

Visibility is part of the story. When the Federal Reserve Banks of Dallas and New York announced a pilot survey in August, they described a U.S. direct lending market estimated at more than $1.3 trillion and noted that visibility into new lending is more limited than in public credit markets. The survey is expected to launch after the third quarter, with aggregate findings anticipated in the first quarter of 2027.¹²

Timing is another open question. The BIS authors note that these loans typically run for several years, a long horizon relative to the pace of change in AI.⁶ And the sources don't yet tell one story: the Congressional Research Service describes AI as already reducing revenue at some software companies, while BIS found no effect yet on the loans themselves. Both can be accurate, because they describe different points in the chain from borrower revenue to loan performance.

The Questions the Data Raises

This isn't a story with a conclusion yet, but it clarifies the questions investors in private credit are asking:

How much of a portfolio depends on a single sector's business model?

Was a loan priced for the risk that exists today, or for the one that existed when it was underwritten?

How do a fund's liquidity terms compare with the liquidity of what it holds?

The question I would ask of any credit exposure is how the assumption behind it could change. Software lending is a case study in how a belief shared across an entire market can become a concentration. The belief here was durable subscription revenue. Whether AI changes it, and how quickly, is what lenders, regulators, and investors are now trying to measure.

Explore more perspectives on private markets, capital allocation, and risk at johnjezzini.com.

Sources

¹ U.S. Congressional Research Service, Private Credit Funds Redemption Restrictions: Market Context and Policy Issues (IN12674), April 2, 2026.

² Bank for International Settlements, BIS Quarterly Review, March 2026, Box B1, as reported in ABF Journal, "Software Lending and the Recurring Revenue Premium."

³ Morningstar, "Why AI Worries About Software Are Hitting Private Credit," 2026.

⁴ PitchBook, "Private credit gets finicky, turning software-sector feast into pecking," 2026.

⁵ Bloomberg, "AI Looms Over Software Companies — and the Investors Who Piled Into Them," 2026 (Fitch ratings data).

⁶ Bank for International Settlements, BIS Bulletin No. 128, "AI disruption in private credit: exposure to software firms in BDCs," July 14, 2026.

⁷ Wellington Management, "What investors should watch as private credit market matures," 2026.

⁸ European Central Bank, Financial Stability Review, May 2026, special feature "Stress in global private credit markets and its implications for euro area financial stability."

⁹ Fortune, "Fed seeks details on U.S. banks' exposure to private credit firms," April 10, 2026.

¹⁰ Semafor, "Private-credit worries spur Fed review," October 5, 2026.

¹¹ Office of Financial Research, U.S. Department of the Treasury, Brief 26-02, "Measuring Counterparty Exposures to Private Credit," March 12, 2026.

¹² Federal Reserve Banks of Dallas and New York, "Federal Reserve Banks of Dallas and New York to Launch Pilot Survey of the Private Credit Market," August 5, 2026.