Anthropic expands legal tools, blocks secondary trades

Anthropic did two very different things on May 12. On one side, it pushed Claude deeper into legal work — a field where “almost right” can be a disaster. On the other, it told investors that a bunch of secondary-market routes into Anthropic stock basically do not count. Put together, the message is pretty clear: Anthropic wants to be more embedded in high-stakes professional workflows, and more restrictive about who gets to touch its cap table. ### What changed in legal? Anthropic expanded Claude for Legal, the law-focused product it introduced earlier this year, with new plug-ins and Model Context Protocol connectors aimed at specific legal tasks and practice areas. The new tools are meant to help with document search, review, deposition prep, drafting, and related clerical-heavy work across commercial, privacy, corporate, employment, product, and AI-governance matters. Anthropic also said the features are available to paying Claude customers. (techcrunch.com) ### Why do connectors matter so much? Because generic chatbot answers are not what law firms are paying for. The real value is getting Claude into the systems lawyers already use — document stores, research databases, workflow tools — so the model can work with actual firm context instead of a blank prompt box. Anthropic’s new connectors tie Claude into tools like Box, DocuSign, and Thomson Reuters products, which makes Claude look less like a standalone chatbot and more like a front door into legal software. (techcrunch.com) ### Why is Thomson Reuters the important tell? That partnership shows where Anthropic thinks the market is going. Thomson Reuters said its new MCP integration lets lawyers move between Claude and CoCounsel Legal, with access to citation-grounded workflows and the company’s legal content stack. Thomson Reuters framed that stack around 1.9 billion Westlaw and Practical Law documents and 1.4 billion KeyCite validity signals — which is a fancy way of saying Anthropic is trying to plug frontier AI into systems that legal buyers already trust. (techcrunch.com) ### So why the stock warning on the same day? Because demand for Anthropic exposure has gotten hot enough that secondary platforms, SPVs, and token-like wrappers are trying to sell it before any IPO. Anthropic updated its website to name Open Doors Partners, Unicorns Exchange, Pachamama Capital, Lionheart Ventures, Hiive’s new offerings, Forge Global’s new offerings, Sydecar, and Upmarket as unauthorized. Anthropic’s position was blunt: any sale or transfer of Anthropic stock, or any interest in that stock, offered through those firms is void and won’t be recognized on its books. (thomsonreuters.com) ### Why is that such a big deal? Because private-company shares are not like public stocks you can just route through any marketplace. Transfers often need board approval and must follow company restrictions. Anthropic is making that explicit at a moment when investors are desperate for AI exposure. One executive from Unicorns Exchange said the platform had fielded more than 50 institutional inquiries for Anthropic shares in three months, with aggregate demand said to exceed $1 trillion — even though none of those introductions closed. (techcrunch.com) ### Is this just about fraud? Partly, but not only. Fraud is the obvious risk — fake certificates, fake ownership claims, messy SPV chains. But the deeper issue is control. Anthropic is still private, and private companies usually want to decide who their shareholders are, how shares move, and whether synthetic or indirect exposure gets treated as real ownership. This warning says Anthropic is not letting market appetite rewrite those rules for it. (techcrunch.com) ### What ties these two stories together? Both moves are about bounded trust. In legal work, Anthropic is trying to win by operating inside professional guardrails — connectors, verified sources, workflow integrations. In private markets, Anthropic is drawing harder guardrails around ownership itself. Same instinct, different arena: expand where precision and control matter, and shut down the parts that get fuzzy fast. (techcrunch.com) ### Bottom line? Anthropic is not just shipping a chatbot anymore. It is picking specific, high-value industries where AI can become infrastructure, while also acting like a company that knows its private shares have become a speculative product of their own. That combination matters — because it shows Anthropic trying to control both the work Claude does and the financial frenzy around the company itself. (techcrunch.com)