Morgan Stanley Warns of Growing Risk in Software Lending: What Asset-Backed Lenders Should Know

Morgan Stanley has flagged growing credit risk in the $235 billion software lending market, warning of deteriorating credit quality, concentrated refinancing risk, and emerging AI disruption concerns that could depress the collateral values backing these loans. The warning is not theoretical: 50% of software loans carry B- or lower ratings, 30% mature by 2028, and 78% are issued to sponsor-backed (leveraged buyout) companies carrying significant acquisition debt.^{1 2}

This matters to alternative business lenders, equipment finance companies, and MCA funders for a reason that has nothing to do with software: the BDCs and private credit funds that hold software debt are the same institutions that provide your warehouse lines, forward-flow arrangements, and working capital facilities. When software defaults spike, those institutions do not tighten selectively. They conduct portfolio-wide reviews that affect every lending vertical they touch.

The Credit Quality Breakdown

The data Morgan Stanley published deserves your attention because of its specificity:

• $235 billion in total software loans, representing 16% of the $1.5 trillion U.S. loan market¹
• 50% of software loans rated B- or lower (high default risk)¹
• 26% rated CCC (speculative grade)¹
• 80%+ issued to private companies with limited financial transparency¹
• 78% are sponsor-backed (leveraged buyout structures with heavy acquisition debt)¹
• 30% of outstanding software loans mature by 2028 (vs. 22% for the overall loan market)¹
• 46% of all software debt comes due within four years (vs. 35% for the broader market)¹

Software accounts for approximately 17% of BDC investments by deal count, second only to commercial services. This is not a niche exposure. It is a structural feature of BDC portfolios.³

KBRA's analysis adds a layer of concern: 25 to 35 percent of private credit portfolios face heightened threats from the AI boom, based on analysis of listed BDC portfolios. The risk extends beyond pure software companies to any business whose value depends on legacy technology products that AI could replace or diminish.^{3 4}

The AI Disruption Risk to Collateral

The core concern is not that software companies will suddenly fail. It is that AI disruption could gradually reduce the value of the recurring revenue streams that back these loans. Software lending uses annual recurring revenue (ARR) as the primary underwriting metric, with lenders structuring facilities around ARR-based leverage multiples. KBRA's research suggests ARR may be a less reliable indicator of credit quality than the market assumes, particularly as AI alternatives reduce customer willingness to pay for legacy software products.³

The collateral recovery problem compounds the risk. Software assets are primarily intangible: intellectual property, customer contracts, code, and brand value. They are not physical assets you can seize and liquidate. In a scenario where a borrower's enterprise value contracts by 40 to 50 percent, recovery rates could fall well below historical norms for software lending because there is nothing physical to sell.⁵

UBS Group has modeled the worst case: in an aggressive AI disruption scenario, default rates in U.S. private credit could climb to 13%, significantly higher than stress projections for leveraged loans and high-yield bonds. The bank expects defaults to increase by roughly 2% in 2026 even in its baseline scenario.⁵

For lenders who extend credit against any form of technology collateral (software licenses, SaaS subscriptions, IT equipment), the warning applies directly. The value of the assets you are lending against may be depreciating faster than your loan term.

Most alternative business lenders are not funding enterprise SaaS companies with $50 million in ARR. They are funding landscaping companies, trucking outfits, and restaurant chains. The direct AI disruption risk to typical MCA borrowers is minimal. The indirect risk, that the BDCs and private credit funds backing your warehouse lines are absorbing software lending losses that tighten capital across all verticals, is the actually compelling concern.

The Maturity Wall

The refinancing risk is concentrated and approaching fast. With 30% of software loans maturing by 2028 and 46% due within four years, there is a $70 to $100 billion refinancing wall ahead. Companies that took on LBO debt when rates were lower and software valuations were higher will need to refinance in an environment where credit quality, valuations, institutional appetite, and interest rates have all moved against them.

The companies that cannot refinance at acceptable terms will restructure, sell, or default. CNBC reported in February 2026 that private credit worries are resurfacing specifically because AI is pressuring software firm valuations. PitchBook noted that investors are souring on software as a lending category, with private credit loans in the sector coming "into sharp relief."^{6 7}

The Contagion Path to Alternative Lending

Here is the connection that most MCA funders and equipment finance lenders miss: BDCs do not operate in isolated verticals.

A BDC with 17% of its portfolio in software, 12% in healthcare, and 8% in alternative lending platforms manages all of these as a single credit portfolio. When software defaults create losses, the portfolio manager's response is not "tighten software lending and leave everything else alone." The response is "review all risk exposures, raise underwriting standards across the board, and reduce facility commitments where we can."

This is the same contagion mechanism we described in our analysis of the non-bank lending boom starting to crack. Private credit CRE stress and software lending stress are additive: both draw from the same BDC capital pool, and both are showing deteriorating credit quality simultaneously. For alternative lenders whose funding comes from these institutions, the capital market headwinds are building from multiple directions.

Bank loan commitments to BDCs have more than doubled since 2021, surpassing $60 billion. Banks charged BDCs higher rate premiums, which BDCs passed to their borrowers. Public BDCs now receive an average of 8% of investment income via payment-in-kind (PIK), meaning borrowers are increasingly paying with additional debt rather than cash. When PIK rises across a BDC portfolio, it signals systemic borrower stress, not just isolated problems.⁸

Corporate bankruptcies are already running at decade highs, adding another pressure point for BDC portfolios that span multiple lending categories. See: 371 Corp Bankruptcies Hit Decade High.

What This Means for Borrower Verification

The software lending risk highlights a broader verification problem: lenders are underwriting against metrics (ARR, recurring revenue, MRR) that may not reflect the borrower's actual operational reality. For lenders extending credit to technology companies, the verification challenge is that financial metrics (ARR, MRR) may be lagging indicators. The revenue figure on today's application may not predict next year's reality. Underwriting that relies solely on revenue data without confirming the borrower's operational fundamentals leaves a gap that widens as AI alternatives accelerate.

The Viola Credit $2 billion ABL fund closing demonstrates that institutional capital is still flowing into asset-backed lending, but increasingly with conditions attached around verification, collateral monitoring, and portfolio reporting. Lenders who can demonstrate institutional-grade underwriting processes will maintain access to capital. Those who cannot will face tighter terms or facility reductions.

See: Viola Credit Closes $2 Billion ABL Fund.

Three Actions for Monday Morning

1. Check your warehouse lender's software exposure. If your credit facility comes from a BDC or private credit fund, ask directly: what percentage of their portfolio is in software? What is the default trend? What is the PIK percentage? You have a right to understand the risk profile of the institution funding your operations. If their software book deteriorates, your facility terms may change.
2. Audit your technology borrower portfolio. If you are extending MCA, revenue-based financing, or equipment finance to SaaS or software companies, review the concentration. What percentage of your portfolio is in technology? Are those borrowers SaaS companies with ARR-dependent models? If AI disruption compresses their revenue, can they still service your advance? Model the impact of a 20 to 30% revenue decline across your tech borrowers.
3. Use the 50% B-or-lower statistic as a benchmark. In your next risk committee meeting, present Morgan Stanley's finding that half of all software loans are junk-rated. Then ask the question: what is the quality distribution in our own portfolio? If you do not know the answer, you have a data problem that needs solving before capital conditions tighten further.

Schedule a free demo to see how Cobalt Intelligence's SOS API adds entity verification to your underwriting process.