How private credit investors are preparing for software’s AI reckoning

Money is pouring into the sector, with assets under management forecast to exceed $2.6 trillion by 2029. Long-term median net internal rates of return were approximately 10.1 per cent across 2010-2021 vintages. Direct lending defaults fell to 1.5 per cent last year.

Yet beneath the surface, a structural risk is building that many private credit investors have not fully sized: a high concentration of loans to software companies whose business models are being torpedoed by AI.

“A lot of these software companies are not going away carte blanche, but their economic models can definitely change and will change,” says Blair Faulstich, senior managing director, head of US direct lending, and portfolio manager with Benefit Street Partners (BSP), Franklin Templeton’s dedicated alternative credit manager.

“The real moment of truth for this is going to be around the refinancing cycle. In 2027 and ’28, there’s a significant software maturity wall. At that point, it’s going to be really interesting to see how it plays out.”

The software exposure gap

Some of the biggest players in private credit including Blue Owl, KKR, Ares, Apollo and BlackRock posted significant outflows from retail private credit funds in the first half of the year as sentiment turned on rising software sector concerns and the US-Iran conflict.

Software companies now account for more than 20 per cent of the BDC industry’s assets, while the exposure can exceed 50 per cent of net asset value in leveraged private credit vehicles.

Several funds enforced a 5 per cent gate to slow redemptions from the semi-liquid vehicles.

“If people had to sell these assets today, people like me would buy them at deep discounts – below fair market value,” says Faulstich, who manages money primarily from institutional investors. “I actually see the proration as protection for retail investors.”

Faulstich says selected deals made before rates rose post-Covid are also under pressure.

It is one of the few areas that Local Pensions Partnership Investments (LPPI) investment director for credit and fixed income, Andrew John, says is showing some stress, although overall LPPI’s overall portfolio fundamentals remain healthy.

“The main area where we have seen some pressure is in 2021 and 2022 vintage loans, particularly deals originated at slightly higher leverage levels before the sharp rise in interest rates,” he says. “We have seen an, albeit manageable, increase in watchlist names from that cohort across the market.”

He says LPPI is still seeing solid EBITDA growth and interest coverage ratios, while defaults and credit losses remain broadly within historical ranges. Its private credit investments span corporate direct lending, asset-backed lending, real estate debt and opportunistic credit.

BSP’s own software exposure sits at approximately 9 per cent, against an industry average of 22 to 24 per cent. Faulstich attributes that to a decision in 2018-19 to pull back from the sector, well before AI entered the picture.

“We were uncomfortable with leverage levels and the credit documents. With software, the asset is your IP and a lot of documents were playing games with the ability to extract that value away from the lenders.”

Jennifer Shum, senior managing director, structured and private credit at the Healthcare of Ontario Pension Plan (HOOPP), argues no-one can claim a clean sheet on software exposure.

“Software is such an important sector – everyone uses software. Everyone will have somewhere between five-plus per cent exposure, while the rule of thumb is 25 per cent is maybe a little bit high.”

HOOPP, she says, sits inside that band and, after reviewing every software holding, is comfortable with its exposure. The broader problem for the industry is what counts as software in the first place.

“They’re effectively software, but they’re not calling it that. Just because I have a subscription doesn’t mean that I’m going to pay for it forever… people treated software exposure like it was a permanent subscription.”

Where the illiquidity premium still exists

While the private credit market has become more challenging since 2024, overall conditions remain attractive, says John.

“In corporate direct lending specifically, spreads have tightened over the last couple of years. That said, this year the US corporate lending market has become somewhat more attractive for new origination, as redemptions have helped rebalance supply and demand.”

For Shum, the retreat of retail money from the private credit sector is doing more than rebalancing pricing. It is handing lenders back the upper hand on terms.

“When retail was buying private credit through the BDCs, the demand outstripped the supply. And you saw that in spreads and in covenants. I was very unhappy that covenants were looser,” she says. “When retail isn’t in, I get my covenants back. Now I actually have some leverage and I can protect our investments the way I’d like to.”

That leverage is greatest in the middle market, where competition never compressed terms as far in the first place.

“There’s lesser demand in that middle market area, so that was always a better sweet spot for covenants and spreads,” Shum says. “Because we didn’t see as much compression in the middle market, you’re not going to see that loosen up as much either.”

Faulstich says growing competition in direct lending to the upper middle market (above $1 billion) has intensified. The result is eroding maintenance covenants, where fewer than half include ongoing financial tests. Meanwhile, more than one-third of direct lending deals above $750 million now include payment-in-kind features.

Given the BSL market wants to execute those megadeals, it leaves sponsors with no reason to pay a meaningful spread premium to private credit lenders.

“The large multi-billion dollar is a great asset-gathering model and they spin it as big companies, safer companies. But I would view it as illiquid risk with modest illiquidity premium. Why would you lock yourself into it if you’re not earning an appropriate illiquidity premium?”

The core middle market – businesses with US$30 to US$75 million in earnings, typically financed at around US$250 million – is where a genuine premium persists.

“The BSL market typically can’t do that deal,” Faulstich says. “So you can earn an illiquidity premium in that part of the market because there isn’t really an alternative.”

LPPI has generally preferred the core middle market where lender protections and overall terms have been more attractive than the upper middle market. Nest, the UK’s largest workplace pension fund, is another asset owner that prefers the middle market alongside its significant asset-backed lending to the real estate and infrastructure sectors.

“We like the middle market segment because it tends to have better underwriting, better contracts and more direct relationships with the borrower,” Nest’s director of public and private markets, Rachel Farrell, explains the fund’s rationale in similar terms, recently told Top1000funds.com.

Beyond direct lending: where allocators are moving

While the core middle market remains attractive, there is a broader shift among institutional investors toward diversification across sub-strategies. More than half (51 per cent) of investors surveyed by BSP plan to increase their overall alternative credit allocation over the next 12 months, with infrastructure debt, asset-based lending and direct lending attracting the largest share of new commitments.

LPPI has been active on several fronts.

“Recently, we have made additional commitments to real estate debt, where rebased valuations following higher interest rates have created opportunities,” John says. “We also see attractive opportunities in private credit secondaries, where investors can buy highly diversified loan portfolios at discounts to fair value.”

HOOPP has tilted the same way.

“Real estate – that’s been really beaten up,” she says. “Many people thought office was dead. It’s roared back. It’s hard to find really great office space these days. So we were a little contrarian on that.”

It is also finding opportunities that sit one step back from the AI boom rather than the software riding on top of it.

“We made an investment in a company recently that produces electronic components for the energy grid that is powering AI,” Shum says. “We want the physical assets that power the AI and that are mission critical. We like those deals and we’ve invested in those deals.”

For BSP, the opportunity set has tilted toward sectors with minimal AI disruption risk – industrials, business services and other non-technology businesses.

“I always tell my team, we’re in the illiquid investing business,” Faulstich says. “You have to tell me, not today, but in year three, what does the business model look like? If you don’t know the answer, we can’t do that deal.”

However, competition remains strong, according to Shum,

“I think because software has been such a talked-about sector and the demand is less, that means that anything else that isn’t software, that demand is still high. Those deals are still the deals that we want. Unfortunately, they are not widening in pricing as the software deals are because demand for software is lower.

“But that just means that, again, things around software may present more opportunities as long as we are very disciplined, continue to look for good covenants and higher pricing.”

For all the warnings about greater regulation, and a surge in retail private credit outflows, Shum sees no cause yet for alarm.

“I don’t see any players currently in the market where it’s going to spin itself out of control. We’re not at the point where there is so much excess in private credit that I feel like that self-regulation is not adequate. When I reflect on AI, I feel like there’s much more need for regulation there than there is for regulation in private credit.”