Private credit defaults fall despite AI software panic

The fear in private credit right now is concentrated in one word: software, and analysts at S&P Global Ratings have spent the past year fielding questions about whether artificial intelligence will hollow out the sector that makes up a large slice of direct-lending books. The data, at least so far, tells a quieter story: S&P's US private credit default rate fell to 3.9% at the end of June 2026, down from 5.3% at the start of the year, according to figures shared with Alternative Credit Investor. The gap between anxiety and actual credit performance is the most interesting thing in this report.

By comparison, the US speculative-grade default rate sat at 3.7% at the end of 2025 and stayed there through June, meaning private credit defaults actually moved down this year while high-yield bonds were flat — hardly the trajectory one would expect from a market supposedly bracing for an AI-led wipeout in software lending.

Evan Gunter, head of private markets research at S&P Global Ratings, told Alternative Credit Investor that AI disruption has been "overhanging the market," and despite supply-side shocks and that uncertainty there has been no matching decline or uptick in defaults — the rate has fallen 140 basis points in six months.

The market prices itself on yield spreads and lender confidence, and if defaults are steady while the narrative is apocalyptic, it is effectively paying managers to sort through the noise. The ones who can distinguish software companies genuinely threatened by AI from those merely marked down by association will be the ones who justify their premium coupons.

None of that makes the environment calm. S&P's chief EMEA economist, Sylvain Broyer, told Alternative Credit Investor that inflation faces a "never-ending succession of supply-side shocks," and the indirect effects of Middle East events on inflation are only now beginning to show. He expects both the European Central Bank and the Federal Reserve to hike rates again later this year, though he thinks long-term rates are near their peak absent another major shock.

Rate hikes push costs up for borrowers on existing floating-rate loans, which can strain servicing, but they also lift yields on new loans, a dynamic that has historically favored the asset class. As this publication argued during the direct-lending reset, the industry is competing on selection rather than volume, and a market that keeps defaults low while macro volatility persists is one where underwriting discipline is being rewarded.

The M&A slowdown compounds that pressure: private credit deal activity has been weighed down by persistent weakness in mergers and acquisitions, even though some participants have reported an uptick in recent weeks. Fewer new loans to underwrite pushes managers to either lower credit standards to maintain deployment or hold more cash, and the S&P default data suggests standards have not yet cracked.

Software is where the AI story and the credit story meet. Private credit carries significant exposure across recurring-revenue loans and buyout facilities for SaaS companies, and the fear is that AI makes some products obsolete or compresses pricing power across the industry. Gunter's comments suggest that fear has not yet turned into realized losses, but the uncertainty itself changes the conversation in underwriting committees.

Managers who thrive will distinguish between software businesses with moats and those with features, not products; the ones who pitch AI exposure as a reason to demand higher spreads are selling the same risk every lender already has, while those who pitch AI due diligence as a reason to select credits are earning their premium. In a market where default rates are steady but the narrative is volatile, the latter group is the one worth backing.

S&P's expectation that the US speculative-grade default rate will stay broadly stable around 4% by March 2027 is a forecast, not a guarantee; it assumes no further major macro shock and that the supply-side disruptions Broyer describes do not spiral into something worse. But it does mean the ratings agency sees no wave of private credit defaults building on the horizon, even with the AI question hanging over the software sector.