Anthropic против LegalTech: как $20 в месяц превратили $300 млрд в пыль

Imagine spending years building a fortress of data, connections, and exclusive access to information, only to have someone walk in and hand out the keys to everyone for the price of a cup of coffee. This is exactly what happened in early February 2026, when Anthropic released its new autonomous agents. The LegalTech market, which once seemed like an unshakeable monolith, collapsed like a house of cards. Shares of giants like Thomson Reuters and Gartner crashed, erasing approximately $300 billion in market capitalization. And if you think this is a temporary correction, you're likely mistaken. This is the beginning of the end for a model where intellectual work is paid by the hour.

For a long time, companies like Thomson Reuters lived off owning the 'pipeline'. They collected legal precedents, analytics, and regulatory documents, packaged them in beautiful interfaces, and sold them to major law firms for outrageous sums. This was a business built on scarcity: there was plenty of information, but only expensive specialists or even more expensive tools could find and correctly interpret what was needed. Anthropic, however, offered not a 'search tool', but a full-fledged 'employee'. Their new agents don't simply search for documents — they understand them, compare them, and deliver ready-made defense or attack strategies.

When investors saw the capabilities of the new models in real time, a domino effect occurred. It turned out that an agent costing $20 a month could handle contract analysis and memo preparation faster than an entire department of junior lawyers. The stock market panic was caused not simply by fear of new technology, but by the realization that the old business model of 'selling access' no longer works. If AI can generate value in seconds, why pay thousands of dollars for a database subscription? Gartner and other analytics giants fell for the same reason: their expert reports now look like yesterday's newspapers compared to the dynamic analysis that neural networks produce.

This collapse is no accident, but the logical outcome of an arms race in Silicon Valley. Anthropic, which has always positioned itself as a 'safe' alternative to OpenAI, this time decided to strike at the most vulnerable spot of the corporate world — efficiency. While competitors argued about how to make a chatbot more human, Dario Amodei's team focused on autonomy. Their agents are capable of independently switching between browser tabs, working with Excel spreadsheets, and even sending emails on behalf of the user. For a legal department of a large company, this means 80% of routine work can be automated tomorrow.

Of course, Thomson Reuters representatives are trying to put on a brave face. They talk about 'unique data' and 'human control', but the market no longer believes them. The problem is that the 'uniqueness' of data rapidly depreciates when AI is trained on the entire body of human knowledge. We are entering an era where value lies not in owning information, but in the speed of its processing and the quality of decisions made based on it. And in this new reality, massive corporations with their bureaucracy and sluggishness lose out to flexible algorithms.

What does this mean for everyone else? LegalTech is just the beginning. Next are financial consulting, auditing, and even classical management. If your business is built on reselling others' expertise or providing access to closed archives, you have problems. Investors will now look not at the size of your database, but at how deeply AI is integrated into your processes. And if you cannot offer something that an Anthropic agent won't do for $20, your place in the unicorn graveyard is already booked.

The bottom line: The old guard of LegalTech missed the moment when AI turned from helper to competitor. Can they adapt, or is $300 billion just the first installment in payment for technological progress?