by Suraj Malik - 1 day ago - 5 min read
A new set of AI tools from Anthropic has triggered one of the sharpest selloffs in software stocks in recent years, forcing investors to confront an uncomfortable question. If AI agents can perform end-to-end knowledge work, what happens to the business models of traditional SaaS companies?
Following CNBC’s reporting, markets reacted swiftly after Anthropic unveiled expanded Claude Cowork capabilities. Within days, hundreds of billions of dollars in market value disappeared across software, data, and IT services companies as fears spread that AI agents could undermine the foundations of subscription-based enterprise software.
Anthropic introduced new Claude Cowork tools that allow AI agents to automate complex, multi-step workflows across fields such as legal research, sales operations, marketing analysis, contract review, and reporting.
Unlike narrow AI features embedded inside existing software, these tools position the AI as the primary interface. The agent can pull data from multiple systems, perform reasoning and analysis, update records, and generate outputs without users needing to log into individual applications.
To investors, this looked less like a productivity enhancement and more like a potential replacement layer. If an AI agent can sit above CRM systems, research platforms, and analytics tools, then those products risk becoming interchangeable back-end utilities rather than differentiated businesses.
The selloff was immediate and broad.
A Goldman Sachs index tracking US software stocks fell about six percent in a single session, its steepest one-day decline in months. Bloomberg estimates that roughly $285 billion in market capitalization vanished across software, financial services, and asset management stocks in one day alone.
Over the course of several trading sessions, the S&P software and services index dropped nearly thirteen percent and now sits more than twenty-five percent below its October peak, even as the broader market continued to hit record highs.
The damage extended across categories. Legal and data providers such as Thomson Reuters, RELX, Wolters Kluwer, S&P Global, and FactSet suffered double-digit declines. Major SaaS names including Salesforce, Adobe, Intuit, and CrowdStrike fell between two and seven percent.
The shockwaves reached global IT services firms as well. US-listed shares of Infosys and Wipro slid sharply, while Accenture and Cognizant lost close to ten percent as clients reassessed long-term demand for traditional outsourcing and managed services.
Even private credit firms exposed to software financing were not spared. Blue Owl, Ares, and KKR posted steep single-day losses as investors priced in higher risk across the sector.
The concern is not one product or one company. It is about where value sits in the software stack.
For years, SaaS companies have justified recurring fees by packaging workflows, analytics, and domain-specific interfaces into tightly integrated products. AI agents threaten to rearrange that structure.
In an agent-first world, users interact with a single intelligent layer. That layer calls multiple APIs, pulls data from different vendors, performs automation, and delivers results directly. The underlying software becomes a data source rather than the destination.
This shift raises several fears.
First, pricing power could weaken if AI agents make it easier to substitute one SaaS provider for another. Second, switching costs could fall if agents handle integrations dynamically rather than locking users into a single platform. Third, valuation models built on multi-year subscription growth become harder to defend when the pace of disruption is unclear.
Not everyone agrees that the selloff reflects a realistic outcome.
Some hedge funds had already built large short positions in software stocks before the announcement, and Anthropic’s launch acted as a catalyst for more aggressive selling. In that view, the market move reflects positioning and momentum as much as fundamentals.
More cautious analysts argue that fears of a sudden collapse in SaaS are overstated. Nvidia CEO Jensen Huang publicly dismissed the idea that AI would eliminate software, saying it would increase demand for software and complexity rather than reduce it.
Research notes cited by CNBC emphasize that enterprises have invested decades and trillions of dollars building software systems and data estates. They are unlikely to abandon that infrastructure overnight in favor of running everything directly through AI platforms.
Instead, many expect AI agents to be absorbed into existing software as automation layers and copilots. In this scenario, some vendors will lose relevance, but others will strengthen their position by adapting quickly.
This is not the first time AI announcements have shaken software valuations. Similar reactions followed OpenAI demonstrations in 2025 and earlier reports warning that dozens of software companies were already under pressure from AI-enabled competition.
What makes this moment different is the scope of the workflows targeted. Legal research, analytics, customer support, and internal operations are among the highest-margin areas of enterprise software. Seeing AI agents operate across those domains concentrated investor anxiety into a single moment.
Anthropic’s tools did not make SaaS obsolete overnight. What they did was force a rapid repricing of expectations.
The market is now grappling with a future where AI agents increasingly mediate how software is used and paid for. In that future, some companies will be relegated to infrastructure roles with lower margins, while others that adapt their platforms around AI orchestration could emerge stronger.
For now, the selloff reflects uncertainty rather than clarity. The software industry is not ending, but its shape is being renegotiated in real time. Investors are reacting not to certainty, but to the realization that the old rules may no longer apply.