Anthropic published a lawyer-related plugin on 30 January, and within four to five days, software stocks had lost billions of dollars. The pace of the collapse left everyone shocked, as the whole industry had been operating on a fundamental misperception of how the AI economy works.
Why Everyone Thought They Were Safe
The narrative that dominated the technological field held that AI companies would build baseline models, which would then serve as a foundation for software companies to design specialized industry-specific tools.
The AI companies would gain through licensing their technology, while the software companies would gain through the sale of completed products to end users, thereby creating a Win/Win ecosystem.
This narrative has attracted substantial capital over the past year.
In December 2025, the valuation of Harvey AI increased 3 times over the previous 12 months, after the company raised $160 million, valuing it at $8 billion, up from $3 billion in February 2025.
Legora, in October, raised $150 million with a valuation of $1.8 billion.
Both companies built on top of Claude by Anthropic (and other top AI companies) via industry-specific interfaces and workflows tailored for law firms.
Two reasons made investors feel that these companies had defendable market positions: 1) they had developed a relationship with prominent law firms; 2) they were aware of the needs of legal professionals and how to integrate with established legal software systems. The engine was simply the basic model; the real value seemed to be all that had been engineered around the basic model.
The Foundation Ate the Stack
Anthropic’s January 30 release included plugins for Claude Cowork. Notably, one of these plugins automates legal procedures, including NDAs, regulatory compliance monitoring, and legal briefings. These features are exactly what companies like Harvey and Legora charge their clients thousands of dollars for in order to streamline.
The technical complexity of the plugin did not stand out as particularly impressive, as it is mostly structured prompts and workflow instructions that teach Claude how to handle legal documents.
However, analysts have noted that these plugins are significantly stronger than traditional software because they are agentic; they use specialized sub-agents to execute complex, multi-step legal workflows autonomously.
This simplicity, however, caused market panic. Anthropic has managed to build a competent software assistant with comparatively simple tooling; there is no reason it cannot do the same across all the other verticals in which software companies currently operate as time goes on.
Between the 4th and 5th of February, the market took the poll: Thomson Reuters dropped 19 per cent, RELX dropped 15 per cent, Wolters Kluwer dropped 13 per cent, and LegalZoom dropped 18 per cent, while the Nasdaq dropped 1.4 per cent, and Indian IT-services powerhouses Infosys and Wipro both dropped more than 5 per cent. These were not run-of-the-mill market swings, but an underlying repricing caused by a burst of realization that the premises on which software values were based could be all wrong.
Why Investors Panicked
Morgan Stanley analysts called the development a sign of increasingly fierce competition.
Moreso, there is an underlying worrying trend; if legal software can be disrupted by AI, the same would apply to consulting work and financial services. When companies that were once considered customers of cloud computing are disrupted themselves, the ramifications go well beyond legal technology to the broader technology ecosystem.
The problem is bigger than mere competition, as the software companies felt they were Anthropic’s customers, who would license Claude, develop products based on it, and sell those products to their own customers. Anthropic built the entire workflow and made contact with end users, without using any middlemen.
Anthropic released 11 plugins that included sales, finance, marketing, data analysis, and customer support functionalities, each targeting a specific software category, indicating that Anthropic has the framework to move into any vertical market at its discretion.
The companies, which had assumed they were creating sustainable businesses on top of Claude, have found out they were simply renting space that could be repossessed at any time.
The term “SaaSpocalypse” summarizes the existential danger that software firms have been facing.
Can they form the application layer of the AI economy?
If foundation-model companies can create applications just as effectively, the whole SaaS industry will face questions that have previously been dodged.
Can Software Companies Disappear In One Night?
The CEO of Harvey, Winston Weinberg, sought to offer reassurance, emphasizing that the global legal market is estimated to be worth 1 trillion and that various actors can survive, thereby avoiding a winner-takes-all atmosphere.
This point had some basis on January 30; however, between February 3-4, it was profoundly doubted by investors who realized that it is not only the question of whether the market is large enough to support several companies, but whether software companies can justify their valuation when the foundation-model companies can replicate their core functionality effortlessly.
The stock market has already provided a first reaction, and now the executives are confronted with some awkward questions about differentiation, defensibility, and long-term viability as they attempt to explain why Anthropic, OpenAI, or Google cannot issue a strand of code that makes entire software companies disappear in one night.
