Artificial Intelligence disruption fears weigh on private credit software exposure

Volatility in public software markets driven by generative Artificial Intelligence concerns is spilling into private credit, where software borrowers represent a significant share of business development company and middle market CLO portfolios but currently show broadly solid credit metrics.

Concerns that generative Artificial Intelligence could erode traditional software business models have moved from public equities into public loans and private credit, where software exposure is substantial but pricing transparency is limited. The S&P North American Technology Software Index is down 19% from the start of the year, and within the S&P UBS Leveraged Loan Index, loan prices for the software/IT sector have been falling since early January, down nearly 465 basis points since Jan. 6 to an average bid of 90.4 on Feb. 6, 2026. These declines have been steeper than in the broader leveraged loan market and have reinforced a pessimistic tone toward software credits in opaque private markets.

Private equity and direct lenders have heavily targeted software over the past decade, supported by cloud adoption, asset light models and recurring revenues. Companies in the technology and software sectors attracted $675 billion from private equity firms in 2022, up from $100 billion in 2012, and software and adjacent sectors now dominate exposure in business development companies and middle market collateralized loan obligations. Software and adjacent sectors account for 20% of the companies on an actual count basis, and on a dollar weighted basis software representation is about 19% of total assets for middle market CLO transactions and 20% for BDCs. Across roughly 325 credit estimated companies in software, health care technology and IT services, over 80% of the companies in software saw their revenue grow on a year over year basis, median revenue growth was over 10%, EBITDA grew by 28.2% on a median basis, and leverage came down by 3% on a median basis, indicating solid recent fundamentals despite valuation pressure.

Refinancing risk is building for some borrowers as higher rates and potential Artificial Intelligence disruption intersect with a sizeable maturity wall. About 20% of the loans issued by credit estimated companies from software, health care technology and IT services have their maturity in the next two years, while about 40% of the maturities fall in 2030 and beyond, creating both near term and long dated refinancing challenges if business models do not adapt. At the same time, structural protections in BDCs and middle market CLOs provide meaningful cushions: BDC portfolios and MM CLOs are largely invested in first lien, senior secured loans with higher equity capitalization in software deals, and only 11 MM CLO tranche ratings, less than 1% of total MM CLO ratings, have been downgraded over the past six years including during the pandemic. Stress tests indicate resilience even under a scenario where 30% of loans default in each MM CLO with a 35% recovery assumption, as no ‘AAA’ rated MM CLO tranche was downgraded below ‘BBB+’ and 94% of the nondeferrable ‘AA’ rated CLO tranche ratings remained investment grade. The impact of Artificial Intelligence is expected to vary widely by sector and regulation, with less regulated, data shareable businesses facing faster deployment and greater risk of legacy software displacement, while regulated industries with sensitive data adopt more slowly, making the ultimate credit impact highly differentiated across portfolios.

55

Impact Score

Courts clarify discoverability of artificial intelligence generated data in litigation

Courts are beginning to define when data from generative artificial intelligence tools must be preserved and produced in discovery, reinforcing that traditional e-discovery rules still apply. Companies are urged to build defensible, proportional strategies for identifying, preserving, and protecting artificial intelligence related data.

Graphics processing unit definition and industry context

A graphics processing unit is defined as an electronic circuit dedicated to handling graphics and video, situated within a wider semiconductor knowledge base that spans devices, materials, architectures, and design flows. The entry anchors a large ecosystem of related technologies, standards, and companies that shape modern graphics and compute acceleration.

Arm reconsiders its licensing model as artificial intelligence reshapes chip demand

Arm dominates smartphone chip designs without selling any chips itself, but the rise of artificial intelligence is pressuring the company to rethink how it makes money from its architecture. Its long-standing model of upfront licence fees and slim per-chip royalties is under scrutiny as demand surges for more powerful, specialized processors.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.