The question now is not whether the landscape is changing. It is how deeply the old model will be cut.
Within a few weeks, one of Wall Street's most beloved sectors turned into a field of mass liquidations. Software stocks have lost a total of $1.6 trillion in market value in the first two months of 2026, with most of the losses concentrated in a brutal two-week stretch that traders at Jefferies immediately christened the "SaaSpocalypse." This is because investors have begun to anticipate that artificial intelligence will not simply serve as an enhancement tool but as a replacement force. The question now is not whether the landscape is changing, but how deeply the old model will be disrupted.
From "Favorite Child" to Scapegoat
For more than a decade, business software was considered the most predictable bet on the market. Stable subscriptions, low capital requirements, high profit margins, and an almost ideal business model. The per-seat licensing formula, in particular, was a machine of compounding revenue: every new employee at a client company meant another subscription. SaaS growth rates climbed quarter after quarter through 2021, turning companies like Salesforce, Adobe, and ServiceNow into blue-chip staples of institutional portfolios.
Today, that same model is under attack.
The State Street SPDR S&P Software & Services ETF, which tracks about 140 companies in the sector, has lost about 20% since the beginning of the year and nearly 30% from its autumn highs. This is a sharp reversal, at a time when the broader S&P 500 remains close to historic highs. The S&P North American Software Index recorded a 15% drop in January alone, its worst monthly decline since October 2008, during the depths of the global financial crisis.
The picture for individual stocks is even bleaker. Intuit, the parent company of TurboTax, has recorded losses exceeding 34% year to date. Workday follows with a decline of about 38 percent. Salesforce and ServiceNow have each shed approximately 25%–30%, while Adobe's forward price-to-earnings ratio has collapsed from a five-year average of around 30x to just 12x. Even Palantir Technologies, the darling of 2025 that saw its stock soar more than 80% last year on the back of 70% revenue growth, has surrendered roughly 22% since January.
What makes this sell-off particularly striking is that it is occurring against a backdrop of continued strong earnings. Only about 67% of software companies in the S&P 500 have beaten revenue expectations this earnings season, compared with 83% for the broader tech sector, but nearly all have beaten profit estimates. The market, however, is not rewarding fundamentals. It is punishing the future.
The $1.6 trillion in losses is a figure that reflects not just a correction, but a breakdown in confidence.
The Sharp Drop in Valuations
The shift is clearly reflected in valuation ratios. In 2022, companies in the SPDR software index were trading at more than 47 times their estimated earnings for the following year. Today, the multiple has fallen to around 19 times, even lower than the S&P 500 average, which is around 22 times. Software-sector price-to-sales ratios have compressed from 9x to 6x, levels not seen since the mid-2010s.
Atlassian is a prime example: from a valuation exceeding 250 times future earnings in 2021, it has returned to levels close to 22 times. Snowflake, CrowdStrike, and Shopify, despite their growth profiles, trade at price-to-sales multiples of roughly 14x, 23x, and 13x, respectively, figures that look conspicuously high only because Snowflake and CrowdStrike are still running trailing-twelve-month net losses.
The market is not simply withdrawing its enthusiasm. It is radically revising its expectations. As Jason Lemkin, the widely recognized godfather of SaaS, has pointed out, public SaaS growth rates have declined every single quarter since the 2021 peak, every single quarter. Strip out price increases from recent earnings reports, and the organic growth picture looks far worse than the headline numbers suggest. The AI crash narrative, in Lemkin's analysis, simply gave the market permission to finally reprice what the fundamentals had been signaling for three years.
Some analysts believe the sell-off has overshot reality. Wedbush Securities' Dan Ives has warned that Wall Street is pricing in a "doomsday scenario" that is almost certainly exaggerated. Analysts at JPMorgan and Goldman Sachs have also argued that the sell-off has gone too far. But the bearish camp retorts that this is not a valuation correction; it is a structural repricing for a world in which the per-seat model may be terminally impaired.
The Trigger: February 3, 2026
The recent tension was sparked on February 3, 2026, when Anthropic announced specialized plugins for its Claude Cowork agentic desktop application, tools designed to manage complex professional workflows across legal research, customer relationship management, sales, marketing, and data analysis. In other words, functions that are at the core of many software companies' products.
The Claude Cowork platform, launched in January 2026, represented something qualitatively different from a chatbot. Rather than simply responding to individual queries, it was designed to plan, execute, and iterate through multi-step workflows autonomously, interacting with external enterprise systems via the Model Context Protocol. The legal plugin could automate contract review, NDA triage, compliance workflows, legal briefings, and templated responses, all configurable to an organization's specific policies and risk tolerances.
The market impact was immediate and devastating. On that single trading day, approximately $285 billion in software market capitalization was wiped out. Thomson Reuters, whose Westlaw and Practical Law platforms offer similar capabilities and whose Legal Professionals segment generates roughly half of its $1.46 billion in quarterly revenue, plunged as much as 18% over the following three days. RELX, the parent company of LexisNexis, fell 14%. Dutch legal software provider Wolters Kluwer declined 13%. The London Stock Exchange Group dropped more than 8%. Even diversified information companies like Pearson, Sage, and Experian saw losses ranging from 4% to 10%.
This was the first direct push into legal technology by a major large-language-model provider, broadening a disruption narrative that had previously been driven by AI-native startups like Harvey. The significance, as analysts noted, was that Anthropic was shifting from being a model supplier to becoming a competitor at the application layer, the lucrative domain where enterprise software companies have built their pricing power.
Anthropic itself emphasized that all results must be reviewed by licensed attorneys and that the tool does not provide legal advice. Thomson Reuters CEO John Hasker pushed back, arguing that companies like Anthropic are building tools good at automating general-purpose tasks but lack the decades of curated proprietary data, case law, contract archives, and regulatory materials that constitute an unassailable moat. Thomson Reuters reported a 5% rise in quarterly profit and reaffirmed revenue growth targets of 7.5%–8% for 2026, having spent over $200 million on AI upgrades in 2025 with plans to repeat the investment. But the market was not in a mood to hear about moats. The share price continued to fall.
The cascade was global. In Europe, the Stoxx Europe Software and Computer Services index shed more than 5% in a single session. In Asia, Japanese firm TIS plunged nearly 16%, Trend Micro lost over 7%, and India's Nifty IT index dropped 5.8%, with Tata Consultancy Services and Infosys falling 7% and 7.3%, respectively. A UBS basket of European stocks deemed vulnerable to AI disruption fell 4.9%. Wall Street's fear gauge, the CBOE Volatility Index, rose to 21.77, its highest close since November 2025.
The "Vibe Coding" Threat
Adding to the sense of existential crisis is the rapid rise of so-called "vibe coding", the practice of building software primarily through natural-language interaction with AI, without requiring traditional programming skills. The term, coined by OpenAI co-founder Andrej Karpathy in February 2025, became the Collins English Dictionary Word of the Year for 2025 and entered the Merriam-Webster lexicon within months.
The threat to enterprise software companies is not abstract. In early February 2026, two CNBC journalists with no coding experience used Anthropic's Claude Code tool to build a working clone of a Monday.com project-management dashboard in under an hour. The result was a functioning prototype capable of managing personal workflows, built at zero cost, against a company with a $5 billion market cap. Silicon Valley insiders noted that the most exposed software companies are those that "sit on top of the work", tools like Atlassian, Adobe, HubSpot, Zendesk, and Smartsheet, while systems of record like Salesforce, which anchor businesses with deeply integrated enterprise data, may be harder to clone but are not immune.
Piper Sandler downgraded Adobe, Freshworks, and Vertex in response, with analyst Billy Fitzsimmons warning that "seat-compression and vibe coding narratives could set a ceiling on multiples."
The concept of seat compression is perhaps the more insidious threat. As Lemkin has explained the arithmetic: if ten AI agents can do the work of a hundred sales representatives, you don't need a hundred Salesforce seats anymore; you need ten. The software itself doesn't get replaced; the humans who generate its per-seat revenue do. That is a 90% reduction in seat revenue for the same work output, and it threatens to structurally impair the recurring-revenue model that has underpinned software valuations for a generation.
There are, of course, skeptics. Nvidia CEO Jensen Huang called the idea that artificial intelligence would replace software "completely absurd." Arm Holdings' Rene Haas spoke of an overreaction by the market. Alexey Korotich, chief product officer at Monday.com competitor Wrike, pointed out that vibe-coded dashboards are the easy part; data governance, back-end infrastructure, security, and reliability at enterprise scale are another matter entirely. Security researchers have highlighted that roughly 25% of AI-generated code samples contain vulnerabilities, and AI co-authored pull requests show 2.74 times higher rates of security flaws compared to human-written code.
But even the skeptics acknowledge that the ceiling on software multiples may have permanently lowered.
When Fear Spreads to Debt
The turmoil is not limited to the stock market. The software industry has a significant and growing presence in the corporate bond and private credit markets, especially after the wave of leveraged buyouts by private equity funds over the past decade. And here, the numbers are deeply unsettling.
According to Morgan Stanley's credit research, roughly 80% of software companies in the credit markets are private issuers. Software now accounts for approximately 25% of business development company (BDC) portfolios, closely followed by private credit collateralized loan obligations (CLOs), with the leveraged loan market at about 16%. Technology and software borrowers represent approximately 17% of BDC investments by deal count, second only to commercial services.
JPMorgan has estimated that between $40 billion and $150 billion of loans packaged into CLOs fall within sectors most exposed to AI risk, with approximately $130 billion to $150 billion of this debt held by companies rated B- or lower, firms with little margin for error if revenue growth slows. An additional $51 billion in lower-rated software debt matures in 2028, with another $50 billion due in 2029, raising refinancing concerns.
UBS analyst Matthew Mish, in his "2026 Private Credit Outlook" report, estimated that 25%–35% of private credit portfolios face elevated AI disruption risk. In an aggressive disruption scenario, Mish warned, default rates in U.S. private credit could climb to 13%, significantly higher than the stress projected for leveraged loans (around 8%) and high-yield bonds (approximately 4%). His base case projects $75 billion to $120 billion in fresh defaults across leveraged loans and private credit by year-end 2026.
The Bloomberg U.S. Leveraged Loan Index suffered its steepest monthly decline in more than three years in February, led by the slump in software and services. Since mid-January, the prices of software-linked loans have fallen significantly, and ratings agencies like Morningstar DBRS report that downgrades in private tech credit now outnumber upgrades by three to one.
The pressure is directly affecting large alternative investment managers. Apollo Global Management cut its direct lending funds' software exposure almost by half in 2025, from about 20% at the start of the year. Shares of Blue Owl Capital, TPG, Ares Management, and KKR all fell by double-digit percentages on February 3 alone. BlackRock shed 5%. Private equity firms including Arcmont Asset Management and Hayfin Capital Management have reportedly begun hiring consultants to audit their portfolios for businesses vulnerable to AI displacement.
JPMorgan CEO Jamie Dimon warned about private credit's "cockroaches", cautioning that stress in one borrower can signal more hidden trouble. Mark Zandi, chief economist at Moody's Analytics, flagged the rapid growth in AI-related borrowing, mounting leverage, and a lack of transparency as considerable "yellow flags," adding that while the private credit industry can probably absorb current losses, that may not be the case a year from now.
The Shadow of Artificial Intelligence Over the Workforce
The software sell-off does not exist in isolation. It is part of a broader reckoning with AI's impact on white-collar employment, a development that is accelerating far more rapidly than most forecasters had predicted.
Amazon, the world's second-largest private employer in the United States, has cut 30,000 corporate jobs since October 2025, including 16,000 in January 2026 alone. CEO Andy Jassy has been explicit about the connection, writing to employees that as Amazon rolls out more generative AI and autonomous agents, "we will need fewer people doing some of the jobs that are being done today." Amazon spent $125 billion on data centers in 2025, money redirected, in effect, from people to processors.
Block, formerly Square, announced in late February that it is eliminating more than 4,000 jobs, nearly 40% of its workforce, with CEO Jack Dorsey declaring that smaller teams paired with AI tools "fundamentally change what it means to build and run a company." Pinterest cut 15% of its workforce explicitly to pursue an "AI-forward strategy." Dow eliminated 4,500 positions tied to automation and AI. Chegg, the online education platform, slashed 45% of its workforce, citing the "new realities of AI" as traffic was cannibalized by AI tutoring. Meta CEO Mark Zuckerberg stated in February 2026 that the year would mark when "AI starts to dramatically change the way that we work," noting that projects that once required entire teams are now being accomplished by single individuals.
In total, U.S. employers announced 946,000 job cuts in 2025, the highest year-to-date total since 2020, with over 55,000 explicitly attributed to AI, a twelve-fold increase from just two years earlier. Approximately 30,000 additional tech jobs have been cut in the first two months of 2026 alone. Gartner analysts estimate that by the end of 2026, one in five organizations will use AI to eliminate at least half of their management layers.
This is the connection that haunts the software sell-off most deeply. The real threat to enterprise software is not that AI replaces the software itself; it is that AI reduces the headcount that uses the software. Fewer employees mean fewer seats. Fewer seats mean less revenue. The subscription model that made software the market's safest bet becomes a vulnerability when the customer base is structurally shrinking.
As James St. Aubin, chief investment officer at Ocean Park Asset Management, summarized: "The seemingly wide moats of these companies feel a lot more narrow today. My biggest fear is that this is a canary in the coal mine for the labor market."
A Cycle of Fear or a Structural Upheaval?
The current correction may not simply be the result of exaggeration. Investors are wondering whether the high-subscription model can survive in a world where automation is becoming cheaper and more powerful.
Large companies have invested billions in existing infrastructure and are not going to abandon it overnight. Enterprise systems of record carry deep integration and high switching costs that AI cannot easily replicate. Microsoft, the most prominent example, survived the mobile disruption panic of the 2010s; its stock has risen 789% over the past decade. Historical precedent shows that SaaS panics in 2016 and 2022 both recovered within months.
However, several factors make this moment different. Hyperscalers, Amazon, Google, Microsoft, and Meta, plan to spend $660–$690 billion on AI infrastructure in 2026, nearly doubling 2025 levels, with much of that investment redirected from enterprise software budgets. AI model capabilities are accelerating at a pace that continues to surprise even experts; according to METR, a third-party benchmarking organization, the complexity of tasks that AI models can perform autonomously roughly doubles every seven months, and recent models have broken that trend sharply to the upside. Anthropic's move from model supplier to application-layer competitor signals that the era in which AI companies and software companies were allies may be ending.
At the same time, AI may gradually shift value from traditional platforms to new tools that operate more flexibly and at lower cost. The old SaaS pricing model, per-seat, with regular 10%–20% hikes during contract renewals, is facing calls for transformation. Some companies, including HP, are already exploring outcome-based pricing models that tie fees to measurable results rather than headcount. If this model takes hold, it could reshape revenue economics across the industry.
As usual, the market is not waiting for proof. It is anticipating the possibility.
The Stakes
The $1.6 trillion "slaughter" is the price of uncertainty in an era of technological acceleration.
Artificial intelligence has not yet "killed" software. But it has managed to kill the sense of certainty that surrounded it, the comfortable assumption that recurring subscription revenue would grow forever, that per-seat pricing would compound indefinitely, that the moats were too deep to breach.
The debate between those who see a cyclical panic and those who see a structural inflection has never been more polarized. The optimists point to valuations at multi-year lows, strong fundamentals, and the immense difficulty of replicating enterprise-grade systems. The pessimists point to three years of declining organic growth, a workforce that is being systematically reduced, and an AI capability curve that is steepening, not flattening.
The most telling commentary may have come from Dave Favuzza at Jefferies: "The draconian view is that software will be the next print media or department stores. The fact that the pendulum has swung so far to the sell-everything side suggests there will be super-attractive opportunities. But when I look out to 2026 or 2027, it is hard to see the upside. If Microsoft is struggling, imagine how bad it could be for companies more in the path of disruption."
And in the markets, the loss of certainty is often more painful than the loss of profits.