Anthropic’s AI Tools for Lawyers Spark Major Legal-Tech Sell-Off

In today’s rapidly evolving technological landscape, breakthroughs in artificial intelligence (AI) often trigger significant shifts in the market. Businesses and investors are scurrying to adapt or protect themselves as new advancements threaten established industries.

This trend played out again recently when Anthropic debuted its productivity tools aimed at the legal sector. Following this announcement, data providers and legal software companies saw their share prices tumble as investors reacted with skepticism towards software firms amid rising AI capabilities. According to The Wall Street Journal, two State Street SPDR exchange-traded funds, which track stocks in software, services, and financial data, lost a staggering US$300.6 billion in market value.

The situation mirrors last year’s reaction when the Chinese company DeepSeek unveiled affordable, efficient AI models, causing a significant decline in the stock prices of companies involved in building resource-intensive data centers—a reaction that was ultimately deemed excessive.

Whether the current market reactions will prove similarly overblown remains uncertain, as the AI sector is fraught with unpredictability regarding which companies will emerge victorious. For instance, hardware manufacturers like Celestica Inc. have thrived over the past year by supplying equipment for AI data centers, as demand for their products has surged. In contrast, software developers face a tougher road. Toronto-based Constellation Software Inc., for instance, has seen its stock price decline by over 50 percent in the last year, largely due to fears that AI-driven coding tools could lower the barrier for competitors and empower existing customers to build their own solutions.

The market’s recent reaction also underscores a growing friction between leading AI developers, such as Anthropic, OpenAI, and Google, which create versatile large language models (LLMs), and the companies that utilize these models in their offerings.

Anthropic’s new plug-in for its generative AI platform, designed to assist with legal tasks like document review and risk assessment, poses a challenge not only to well-entrenched legal tech firms but also to numerous AI startups focused on providing tailored tools for the legal industry.

Shares in Thomson Reuters Corp., which is making substantial investments in AI, plummeted by 16 percent on the Toronto Stock Exchange following the announcement. Similarly, CS Disco Inc., an AI-focused legal services provider in the U.S., saw a 12 percent drop. Other companies such as RELX, which owns LexisNexis, and Wolters Kluwer experienced declines of 15 percent and 13 percent, respectively.

A pressing question arises: What might deter Anthropic and similar companies from delving further into the legal technology arena—or other sectors—and crafting bespoke software that could threaten their own client base?

“These applications merely act as a wrapper for existing capabilities embedded in an LLM, and I struggle to see a sustainable competitive advantage here,” remarked John Ruffolo, founder of Maverix Private Equity. “Building your business atop another that could rival you creates inherent risks.” This perspective frames part of Maverix’s reluctance to invest in AI startups despite their soaring valuations.

However, in the legal domain, some experts argue that AI developers may be hesitant to pursue this market aggressively. Scott Stevenson, co-founder of AI legal software firm Spellbook, notes that the financial prospects might not justify the effort, as some specialized services are challenging to replicate. Spellbook, for instance, caters to approximately 4,000 clients utilizing its AI tools for contract analysis.

“The model providers may eventually develop light tools for various sectors, but their offerings are relatively basic,” Mr. Stevenson said. “Spellbook is like a toaster—we focus on a single function and excel at it.” With the introduction of each new LLM, Spellbook’s client base has expanded, indicating a growing appetite for specialized tools among law firms. Nevertheless, advancements in general-purpose AI compel all competitors in the field to elevate their standards.

Other industries might be more susceptible to disruption. Companies like Cursor, Replit, and Lovable have enjoyed considerable success due to the popularity of AI coding tools that create apps and websites from simple textual instructions. Yet, these products, relying on LLMs from pioneering developers, face stiff competition as Anthropic’s Claude Code rises in prominence among software developers, alongside OpenAI’s own coding platform, Codex.

For larger developers, creating coding tools may be easier than navigating intricate integrations with diverse software and processes in various sectors. Additionally, the lucrative nature of the software market provides further incentive for continuous improvement. “I anticipate they will certainly compete with Cursor and other similar entities,” Mr. Stevenson added.

For investors, discernibly identifying which sectors and businesses will be affected by AI remains a significant challenge. The S&P North American Technology Software Index has seen a decline of approximately 19 percent this year. Share prices in video game companies like Take-Two Interactive Software Inc. and Roblox Corp. plummeted recently, following the release of Google’s Project Genie, which can create digital worlds based on text prompts.

“The competition in AI may be the most cutthroat we have ever witnessed in the tech industry,” observed Mr. Stevenson. “Currently, the only sustainable advantage lies in speed.”