
Private equity giant Apollo warns: "Software is dead"! AI is reshaping the $440 billion valuation logic

The long-standing "safe haven" of private equity in the software industry has collapsed due to the impact of AI. Institutions like Apollo have significantly reduced their software exposure, leading to a sharp decline in the valuation multiples of SaaS companies. Credit institutions are urgently reviewing vulnerable assets in their portfolios. This structural disruption triggered by AI directly impacts the $440 billion software investments made by private equity over the past decade, and the old paradigm of "stable growth and high valuation" in the industry is rapidly being dismantled
John Zito of Apollo recently made waves at an investor gathering in Toronto with his provocative question—“Is software dead?” This marks a complete break from the long-held core assumption in the private equity market that the software industry offers “stable growth and substantial revenue.” The rise of artificial intelligence is forcing investors to reassess this once-coveted sector.
This warning is not unfounded. According to insiders who spoke to Bloomberg, Apollo has reduced its direct lending fund's software risk exposure by nearly half by 2025, down significantly from about 20% at the beginning of the year. Meanwhile, institutions like Arcmont Asset Management and Haifen Capital Management have hired consultants to specifically identify companies in their portfolios that may become vulnerable due to the impact of AI technology.
Market panic has begun to spread. Concerns over disappointing returns on AI investments have led to a decline in Microsoft's stock price; a technology fund under Blue Owl Capital has experienced massive capital outflows; and two European software companies have been forced to shelve loan transactions. The traditional Software as a Service (SaaS) model is facing a downward pressure from AI-native companies and “ambient coding” startups, putting once-stable moats in jeopardy.
This upheaval directly impacts the $440 billion bets made in the private equity industry over the past decade. According to Bloomberg data, from 2015 to 2025, private equity buyers spent over $440 billion acquiring more than 1,900 software companies. Now, with rising borrowing costs and accelerated technological iterations, investors and lenders are fundamentally questioning the survival and valuation logic of these assets.
The Collapse of Valuation Systems and the SaaS Crisis
For a long time, software transactions were easily approved by investment committees due to their “sticky” revenue and subscription models. However, this logic is being rewritten by AI. Isaac Kim, a partner at venture capital firm Lightspeed and former head of technology private equity at Elliott Investment Management, bluntly stated, “The current form of tech private equity is dead.” He pointed out that AI has changed the basic assumption that “underlying products can remain relevant long enough for financial engineering operations.”
The SaaS model is the first to bear the brunt. Tools like Claude Code launched by Anthropic allow users without programming experience to build software, significantly lowering the technical barrier and undermining the rigid advantages of traditional SaaS products.
This impact is already reflected in valuations. According to Pitchbook data, the average multiple for private equity acquisitions of SaaS companies has dropped from 24 times the previous year to 18 times in 2025, while some star companies like Coupa Software and Cloudera previously commanded multiples as high as 60 times. Robin Doumar, founder of Park Square, stated that the software industry’s “invulnerable halo” is long outdated, and hopes are that the era of high multiples that defy financial logic will come to an end
The Dilemma of the Credit Market and Default Risks
For borrowers in the software industry, risks have transformed into actual losses. In 2025, due to heightened scrutiny from investors regarding the impact of AI, outsourcing companies KronosNet and Foundever found themselves in distress, with their debts currently at non-performing levels. The bond prices of other software companies, including McAfee and ION Platform Investment Group, have also plummeted. The credit division of CVC Capital Partners was forced to take over the call center support business Sabio Group, as its original owner could not find a buyer.
Jon Gray, President of Blackstone, warned that the biggest risk lies in "disruption," akin to the situation faced by Yellow Pages when the internet emerged in the 1990s. Since the software industry operates on a light asset model, lacking physical assets as collateral, lenders face a greater risk of principal loss if a company goes bankrupt.
Major investment banks have issued warnings:
-
UBS: In a scenario of "aggressive AI disruption," the default rate in the private credit market is expected to rise to 12%-13%.
-
Citigroup: Recommended a bearish stance on software loans, believing the sector has limited appeal outside of CLO investors.
-
Morgan Stanley: Suggested shorting total return swaps on loans due to high software exposure.
-
Barclays: Pointed out that the loan value of software BDCs (Business Development Companies) may rise, facing risks of asset quality deterioration.
Hidden Exposure and Market Overreaction?
It is noteworthy that the actual exposure of private credit to software may be far higher than the reported figures. Raymond James analyst Robert Dodd pointed out that if a software company serves the healthcare industry, funds often classify it as healthcare exposure, meaning the actual risk is underestimated.
Despite the pessimistic market sentiment, the S&P North American Software Index fell 15% in January, marking the largest monthly decline since 2008, but not all voices are bearish. Some cases previously viewed as risky are seeking survival through transformation. For instance, Zendesk did not mention AI risks when it was acquired in 2022, but quickly integrated technology, and its internal AI product's annual recurring revenue has now exceeded $200 million.
Brian Ruder, Co-CEO of Permira, believes the current concerns are somewhat exaggerated. He stated, looking back at historical technology platform shifts, there will be winners and losers among both AI-native companies and existing SaaS giants. However, for lending institutions, the top priority now is to question software company owners in lending meetings: how will you respond to the challenges posed by new AI technologies?
