The Wall Street Journal just reported that about $140 billion of risky corporate loans owed by private-equity-backed companies come due in 2028, with another $160 billion in 2029. From 2023 to 2025, the amount due two years out averaged about $60 billion, per PitchBook.
Software accounts for about one-third of those maturities, or roughly $100 billion of loans against B2B companies that PE firms bought at peak multiples. Many of them are growing slower than their debt assumed, and AI-native competitors are taking share of their customers’ budgets.
So far the software credit cycle has produced one major lender takeover (Medallia) and a large number of quiet workouts. Most of the maturities are still ahead, and the refinancing decisions on them get made in 2027.
#1. Software Is a Third of the 2028–2029 Wall, and 21% of Software Loans Mature in 2028
PE funds bought a lot of software over the last decade and financed those deals with leverage, which is why software is so overrepresented in the 2028 and 2029 maturities.
Morningstar’s data shows how concentrated it is: 21% of outstanding loans in the sector mature in 2028, compared with 14% on index loans broadly.
The syndicated loan numbers also understate total exposure, because private credit holds a large share of software debt on top of them. The BIS found outstanding private credit loans to SaaS firms increased from almost $8 billion in 2015 to over $500 billion, or 19% of total direct loans, by end-2025. Those loans are marked much less transparently than syndicated debt.
#2. Seven-Year Loans Written in 2021 Mature in 2028
Most of this wall dates to 2021. Rates were near zero, PE firms did record megadeals (including the $11.8 billion McAfee takeover), and much of the financing came as seven-year loans because sponsors planned to sell within that window. A seven-year loan written in 2021 matures in 2028.
Those deals assumed rates would stay low and the companies would grow into their capital structures. The Fed started hiking in 2022, Treasury yields rose sharply again this year, and AI pushed software multiples down just as these companies needed to exit.
Medallia shows how this plays out. Thoma Bravo reached an agreement to hand Medallia to its lenders, wiping out $5.1 billion in equity on a $6.4 billion 2021 take-private, with Blackstone, KKR, Apollo and Antares taking control in a debt-for-equity swap. In March, Thoma Bravo said it had overestimated the company’s growth prospects and paid too much for it. The transaction closed on August 3, 2026.
Thoma Bravo is widely considered the strongest operator in software private equity, and it still lost nearly all of its equity on a 2021 deal. When a sponsor of that caliber can’t make a 2021 buyout work, the issue is the entry price and leverage of that vintage, and plenty of other sponsors bought at the same prices.
#3. Software Loans Fell to About 85 Cents While Other Loans Held Near 96
Lenders repriced software debt early this year and left most other sectors alone. Bids for non-software loans largely held around 96 cents on the dollar, while software loans dropped from roughly 95 cents at the end of 2025 to about 85 cents. The selloff began after Anthropic and OpenAI released new models that let individual users build programs that could, at least in theory, replace services many existing software companies sell.
New lending to software has also dried up. Just 9% of all US broadly syndicated loans issued this year, excluding repricings, have come from software companies, after two decades in which software was the market’s largest borrower.
Individual names in the WSJ piece trade well below 85. Solera, the vehicle-maintenance software company Vista acquired in 2016, has about $8 billion due in 2028 and 2029, a CCC-plus rating from S&P, and loans trading between 77 and 93 cents. Some of symplr’s lenders had marked its loans down to about 66 cents.
Loan prices usually fall well before companies actually default, and these prices show lenders expect losses on a meaningful share of this debt.
#4. Moody’s Puts Private Credit Defaults at 4%, or 5.8% Counting PIK and Extensions
Headline default rates look benign. Moody’s reported that leveraged loan defaults fell to 5% in Q2, the lowest level since September 2023.
That figure excludes most of the restructuring activity. Moody’s put the default rate among 607 US private credit borrowers at 4% for the 12 months to June, rising to 5.8% once PIK amendments and maturity extensions involving distressed borrowers are counted as defaults, and software borrowers recorded the weakest credit metrics of any sector.
PIK (payment in kind) lets a company stop paying cash interest and add it to the loan balance instead. The company avoids default this year, and the debt grows. Houlihan Lokey estimates that over a third of private credit agreements with software borrowers at the end of 2025 included a PIK option, a threefold increase in three years. Across the market, around 11% of outstanding private credit loans had some or all interest structured as PIK in Q2, compared with 7% at the end of 2021.
Lenders are now pulling back on PIK. Only 13.5% of new private credit loans originated in Q2 included a PIK provision, down from 25% at the end of 2025, and managers have become more cautious about lending to businesses exposed to AI disruption, particularly software companies. The tools that kept software defaults low in 2026 will be harder to get in 2027.
#5. symplr Got Extended, Sophos Got Turned Down, Medallia Lost the Keys
Recent deals show three different outcomes.
symplr was extended after its sponsor put in more money. Clearlake tried to refinance $2.2 billion of loans for symplr, the healthcare software company it bought in 2018. Ares and Sixth Street refused until Clearlake increased its investment. Clearlake’s continuation fund put in $175 million for preferred shares, and lenders pushed maturities out to 2030 and 2031, added about $100 million in new loans, and let symplr defer some interest.
Sophos was turned down. Sophos, another Thoma Bravo acquisition, approached private credit lenders this year to refinance roughly $2.5 billion in loans coming due next year, and several passed.
Medallia went to its lenders. The new owners are investing $150 million of new capital and the transaction will significantly reduce the company’s debt.
Bain’s Hugh MacArthur told the WSJ that some of these businesses will get “the keys tossed back to the bank.” The symplr route requires a sponsor with fresh capital that still wants to own the company, and many 2021 deals won’t meet either condition.
#6. The Refinancing Window for 2028 Maturities Closes in 2027
Companies typically refinance 18 to 24 months before maturity. For loans due in 2028, that window is open now and closes during 2027. For 2029 maturities, it runs from 2027 into 2028.
Most decisions on the $300 billion will be made over the next 12 to 18 months, after a year in which rates rose, PIK became harder to get, and new software lending shrank to a small share of the market.
Forecasts point the same way. Bank of America said in a report Friday that near-term maturity walls are at their highest level since 2013, and it expects the default rate on below-investment-grade loans to double next year to 4%, with lower recoveries. In Moody’s sample, lower-rated issuers account for more than half of debt maturities due in both 2027 and 2028, leaving them reliant on extensions, PIK amendments, costly refinancing or more support from their PE sponsors. About a quarter of the loans due in 2028 and 2029 already carry ratings of CCC-plus or below.
Plenty of software companies will refinance without trouble. Fitch found that half of the software companies it rates are at low risk of AI disruption, while 9% are at high risk. The defaults will come from companies where moderate AI risk sits on top of high leverage and flat growth, since those companies can’t grow fast enough to support their debt.
#7. What the Debt Wall Means for B2B Founders Without a Term Loan
Most SaaStr readers don’t carry a $2 billion term loan, but this cycle affects them in several concrete ways.
Selling to PE will bring lower prices. With software at 9% of new syndicated loan issuance and lenders cautious about AI exposure, PE buyers can borrow less against your ARR, which means fewer bidders and smaller checks. If you were counting on a sponsor buyout at 2021-era multiples in the next two years, assume a lower price or a strategic buyer.
PE-owned competitors facing 2028 maturities tend to cut R&D and headcount, raise prices on customers who can’t easily switch, and stop funding new products. Make a list of competitors owned by sponsors on 2021-vintage deals, pull their customers’ renewal dates where you can, and target those accounts over the next 18 months.
Companies negotiating with lenders usually lose their strongest engineers and AEs first. If you’re hiring, those teams are worth recruiting from now.
Lender-owned software assets will come up for sale. Blackstone, Apollo and KKR are credit investors and don’t want to run Medallia indefinitely, or the next ten companies that land with them. Expect a steady flow of distressed software companies, product lines and customer bases over the next two years, at prices well below what sponsors paid.
The same math applies to smaller debt. A recurring-revenue loan or venture debt facility assumes ARR keeps growing. If your growth is slowing and your covenants assume otherwise, start renegotiating 18 months before the problem, while you still have options.
The 2028 and 2029 Maturities Are Still Ahead of Us
So far the software credit cycle has meant loan prices down about 10 points, BDC shares down, new software lending down to 9% of issuance, and a wave of PIK toggles, amend-and-extends and continuation-fund rescues. Medallia is the one large case that ended with lenders taking control.
The maturities themselves haven’t arrived. $140 billion comes due in 2028 and $160 billion in 2029, about a third of it software and a quarter of it rated CCC-plus or below, with most of the refinancing work landing in 2027. Rates are higher than when these loans were written, PIK is harder to get, and AI-native vendors are competing for the same budgets these companies depend on. A significant number of them will end up in the same position as Medallia.
