Direct lenders are reassessing exposure to software companies as valuations in the public markets slide and investors debate which businesses will survive a potential AI shakeout.

Publicly traded shares and bonds of software and tech-adjacent companies have sold off sharply in recent weeks as investors adjust their exposure in response to widespread uncertainty in the market about the threat AI poses to traditional business models.

Major alternative asset managers that run private credit strategies likewise have traded down even as senior executives publicly defend their investments in the sector.

While the illiquidity of private credit loans limits funds’ ability to rapidly adjust their portfolios, near-term assessment is more critical than reactive repositioning, said Brett Pearlman, a partner at Cleary Gottlieb who advises price credit investors.

“Private lenders will have important decisions to make with respect to stressed companies in their portfolios,” said Pearlman, “including whether capital stacks should be recut with the sponsors maintaining control, should lenders ‘take the keys,’ or should the companies seek to sell off their assets?”

Vanilla deals are out

Public and private markets are de‑risking from software simultaneously, just in different ways, said a private credit investor.

While private credit retains structural advantages over the BSL market, the investor said funds are being pushed to tighten underwriting and reduce incremental exposure to legacy SaaS models due sector concentration, AI-driven business model risk and, in some cases, redemption pressures from LPs.

Private credit may still show more appetite for idiosyncratic, sponsor-supported situations, the source added, but there is less enthusiasm for broad, vanilla software exposure than during the industry’s heyday of 2021 to 2023.

Liquidity, a private credit lender for late-stage technology companies, is looking to lend to businesses that use AI to “optimize physical processes” and deliver meaningful efficiencies that boost margins, said Chief Strategy Officer Carmen James.

The fund sees the highest risk for disruption in companies with generic models operating in areas like business process outsourcing and legacy SaaS, James added.

The revaluation of software risk is needed after years of exuberance, according to Matt Schwartz, a private credit partner at DLA Piper.

“Software isn’t broken — what’s changed is that investors are underwriting it like a real business again,” said Schwartz. For the first time in years, investors are assessing software companies based on market structure and competitive positioning, rather than metrics based on annual reoccurring revenue.

(Similarly, Alix Partners has forecast that AI will move enterprise software companies’ business models away from predictable per-seat licensing to new pricing mechanisms as customers look to pay based on outcomes and productivity gains.)

Pressure is building on broad, undifferentiated platforms, “that can’t clearly define their market or defend their relevance,” Schwartz said, whereas capital is increasingly rotating toward targeted, AI-enabled software businesses tied to specific workflows and industries, particularly where automation delivers tangible operational value.

Software was ‘priced for perfection’

In a positive for the industry, private credit financed software companies are starting from a position of financial strength as judged by recent earnings, according to an assessment of borrowers monitored by KBRA.

The ratings agency found that software businesses posted the highest EBITDA compound annual growth rate among industries it tracks, alongside the second-lowest proportion of companies with declining EBITDA and the second-highest median EBITDA margin.

“This feels more like a narrative that’s crystallized rather than a sudden shift in fundamentals,” said Jake Mincemoyer, a partner at A&O Shearman, cautioning against treating software as uniquely exposed.

Advisors say it is key to distinguish between AI-related uncertainty and structural issues facing software deals originated during the market peak.

“The pressure on 2021 software deals really reflects a different issue — those deals were underwritten in a very different market and many are facing pressure from higher interest costs than were modelled as well as changes in demand — rather than AI risk alone,” said Mincemoyer. While pricing is higher today, he added that the vintage issue and the AI debate are related but ultimately separate dynamics.

“The real risk today isn’t having software exposure — it’s backing companies that were priced for perfection or that don’t have a credible, practical AI strategy,” added Schwartz.

Impact on new software deals

Market participants are divided on whether the current market rout will impact new deals.

Cleary Gottlieb’s Pearlman said he expects there to be a high bar for regular private credit funds to invest in software names until the market gets a better sense of the extent of AI disruption in the sector.

Mincemoyer, though, said private credit lenders continue to finance new, cash-flowing software businesses, particularly where they are comfortable underwriting fundamentals.

“I’m not seeing a complete risk-off approach to software,” he said. “There are deals where the broadly syndicated market has stepped back, but private credit lenders — having done the work — were comfortable stepping in.”