Software Stocks Post Worst Month Since 2008 as AI Fears Spread to Loan Markets

The S&P North American software index fell 15% in January, its steepest monthly drop since the 2008 financial crisis, Bloomberg reported today. The selloff accelerated this week after Anthropic released a legal productivity plug-in for its Claude chatbot, sending shares of RELX and Wolters Kluwer down more than 10% and dragging the broader iShares Expanded Tech-Software Sector ETF into bear market territory, down 22% from its recent high.

Traders at Jefferies have a name for it. "We call it the 'SaaSpocalypse,'" said Jeffrey Favuzza, who works on the firm's equity trading desk. "Trading is very much 'get me out' style selling."

What makes this rout different from ordinary sector rotation: the fear isn't about a bad quarter or a missed estimate. It's about whether the per-seat subscription model that powered two decades of software investing still works when AI agents can do the job a licensed user used to do. That question started on equity trading desks. Now the contagion is running through credit markets and private equity boardrooms, too, and nobody is sure where it stops.

The numbers that spooked Wall Street

ServiceNow beat fourth-quarter earnings expectations and issued better-than-expected guidance. Announced a buyback. And the stock still fell 10% on Thursday, because none of it mattered. Morgan Stanley's response was blunt. "Good, but not good enough."

That reaction tells you where sentiment sits. In a normal market, ServiceNow's results would have been a positive catalyst. In this one, stable growth "likely falls short of shifting the narrative," the Morgan Stanley analysts wrote in a note to clients.

ServiceNow's forward price-to-earnings multiple has collapsed from the upper 60s in early 2025 to under 28, according to CNBC. The stock is down 49% over the past year. Jim Cramer, who normally champions enterprise software, told his "Mad Money" audience he couldn't fight the selling. "The stock market has said, 'Ain't got nothin for you, Bill,'" he said, referring to CEO Bill McDermott.

Microsoft added to the pressure, falling 10% after reporting slower cloud growth for its fiscal second quarter. January was its worst month in more than a decade. SAP plunged 15% after its cloud contract backlog grew 16%, well below the 26% analysts expected. UBS called it a "disappointment." The German software company, which was Europe's most valuable firm less than a year ago, has shed 40% from its peak.

Piper Sandler downgraded Adobe, Freshworks and Vertex on Monday, warning that "seat-compression and vibe coding narratives could set a ceiling on multiples." Vibe coding, a term for using AI to write software, has become the shorthand investors use when they want to explain why traditional SaaS might not recover.

The contagion hits credit

The panic has jumped asset classes. Software loan prices dropped sharply last week, an unusual divergence from the rest of the leveraged loan market, which surged after President Trump pulled back his tariff threats on Greenland.

A Cloudera loan fell 7 cents on the dollar in a single week, Bloomberg reported. Loans tied to Dayforce, Rocket Software and others also declined. Deutsche Bank's $1.2 billion loan deal for Thoma Bravo's Conga stalled after investors balked at the sector's prospects. European software firm Team.Blue halted a leveraged loan offering entirely.

"A storm has hit the loan market," said Scott Macklin, head of US leveraged finance at Obra Capital. "The heaviest calendar in months has collided with mounting existential questions around software business models."

Software accounts for 12% of the Bloomberg US Leveraged Loan Index, the single largest sector. Nomura data shows software debt in collateralized loan obligations has posted the worst total returns of any sector so far this year. Morgan Stanley published a report last Friday recommending investors short AI-exposed credits and favor junk bonds over leveraged loans because of the latter's heavier software exposure.

Bank of America's Pratik Gupta, who leads CLO and RMBS research, pushed back. "There is a certain element of throwing the baby out with the bathwater," he said. "The software sell-off got pushed into names which likely are not going to be affected by AI."

Private equity's $440 billion question

Nobody captured the mood in buyout circles better than Apollo Global Management's John Zito. Speaking to an audience of investors in Toronto last fall, he posed a question that Bloomberg is now reporting for the first time. "The real risk is, is software dead?"

Private equity firms poured more than $440 billion into over 1,900 software acquisitions between 2015 and 2025, according to Bloomberg data. The thesis was straightforward. Software revenues are sticky, the subscription model produces predictable cash flows, and you can add debt against that predictability. For years, the pitch sailed through investment committees.

AI broke that logic. "Buying a software business, improving margins and adding leverage assumes the underlying product remains relevant long enough for financial engineering to work," Isaac Kim, a partner at Lightspeed who previously ran Elliott Investment Management's tech PE practice, wrote in a LinkedIn post. "AI has changed that assumption."

Apollo cut its direct lending funds' software exposure from about 20% to roughly half that during 2025. Arcmont Asset Management and Hayfin Capital Management hired consultants to audit their portfolios for vulnerable positions. Lenders now ask software executives about AI competition before anything else in borrowing meetings, according to people familiar with the conversations.

The problem gets worse when you try to measure actual exposure. Barclays estimates business development companies hold around 20% software exposure, but Raymond James analyst Robert Dodd says the real number is higher. "If your software business is in healthcare, the fund classifies it as a healthcare exposure," he said. "The software exposure is meaningfully higher than it looks."

Buyout multiples already reflect the shift. SaaS companies sold to private equity at an average of 18 times earnings in 2025, down from 24 the prior year and a far cry from the boom-era peaks near 60 times that companies like Coupa Software once commanded.

The Zendesk parable

There's a story that captures the speed of this transition. In late 2022, Hellman & Friedman and Permira paid just over $10 billion for Zendesk. The company's board filed an SEC document listing the risks it faced going forward: recession, inflation, economic headwinds. AI was not mentioned.

ChatGPT launched eight days after the deal closed.

Zendesk adapted. Its in-house AI offering now generates more than $200 million in annual recurring revenue, about 10% of total revenue, according to a person familiar with the matter. But the broader point stands. An entire investment thesis was built on assumptions that evaporated in a week.

Permira co-CEO Brian Ruder argues the concern has been overdone. "History tells us there will be winners and losers on both sides of the AI-native and incumbent SaaS equation," he told Bloomberg.

Who survives the contagion

The selloff is indiscriminate, but not everyone agrees it should be.

Gil Luria, head of technology research at D.A. Davidson, told CNBC the market is pricing in wholesale destruction that hasn't actually materialized. "AI hasn't actually disrupted any of these businesses, but the conventional wisdom is that it will disrupt all of them," he said. Several software stocks have swung from being labeled AI winners to AI losers and back again, including Salesforce, Adobe and Elastic. William Blair analyst Arjun Bhatia argued in a note to clients last week that fears of AI model disruption "do not justify the broad-based indiscriminate selling we are seeing across the sector."

Beach Point Capital's Sinjin Bowron, who manages liquid credit strategies, made a related point about the debt side. Many of the software companies whose loans are trading down still report solid results and keep gaining customers. "These are incredibly deeply entrenched software suites in company processes," he said. "It would take potentially years to rip out and replace some of these."

Ben Reitzes at Melius Research, who first warned that "AI is eating software" back in April 2024, sees a narrower path to safety. Data infrastructure companies like Snowflake and MongoDB make data usable across systems rather than selling end-user tools that AI can replicate, he argues. If you control the data layer, you become more valuable as AI proliferates. Per-seat application vendors face the opposite math. Reitzes called SaaS incumbents "AI- and agent-last and seat-first" and said their pricing models may not survive.

The Sycomore Sustainable Tech fund, which has beaten 99% of its European peers over three years, bought Microsoft shares during the selloff. At less than 24 times estimated earnings, the stock trades at its cheapest level in about three years. BTIG chief market technician Jonathan Krinsky wrote that the sector is "probably oversold enough for a bounce." But he added a warning. "It is going to take a long time to repair and build a new base."

Blackstone's Jon Gray offered the starkest framing on Bloomberg TV. "I think the biggest risk is actually the disruption risk," he said. "What happens when industries change overnight, like what we saw to the Yellow Pages back in the nineties when the Internet came along."

Favuzza at Jefferies put it more directly. He asks clients where their "hold-your-nose level" is for buying software names. Nobody has an answer. With earnings season still running and the next wave of AI product launches from Anthropic, Google and OpenAI expected in the coming weeks, the contagion has room to spread.

OpenAI's Code Red Produces a Point Release: What GPT-5.2 Reveals About the AI Arms Race

Sam Altman declared a code red. Engineers got their holiday plans cancelled. Product teams that had been working on advertising integrations found themselves reassigned to ChatGPT core. And what emerg

The Implicator

Vibe coding usage crashes months after platform push

Google launched its vibe coding studio a day ago while Chamath Palihapitiya posted charts showing vibe coding usage had already peaked and crashed. The mismatch wasn't subtle—platforms doubled down on

The Implicator

Werner Walks Out. The Hackers Never Left.

Good Morning from San Francisco, Werner Vogels printed sixty thousand newspapers and walked off the Re:Invent stage for the last time. Fourteen years of keynotes. Two words to close it all: "Werner o

The Implicator