The law has never been a profession in a hurry to get ahead of technology; it bills in six-minute increments and still charges clients for photocopies. Then, in just the past year, the dam broke. For example:
Kirkland & Ellis, the world's highest-grossing firm, has committed $500 million to build its own AI platform, enlisting 250 of its lawyers, 180 engineers, and Palantir as a partner.
Morgan & Morgan, the country's largest plaintiffs' firm, recently pledged $1 billion to build its own bespoke AI platform.
Latham & Watkins quietly spent tens of millions buying Nvidia H200 servers and standing up private infrastructure to fine-tune open-weight models on client data too sensitive to hand to any cloud vendor.
Similar efforts and investments are underway across the Top 20 firms in the US.
Three things make this remarkable.
This is a first for big law. Partnerships are built to distribute profit, not to build infrastructure. Historically, the choice between building out a serious Salesforce or Workday solution Vs. the year-end vacation to Aruba was no choice at all. AI has flipped the math, and Big Law is now investing more aggressively than most industries. That tells you how squarely AI represents both opportunity and threat in knowledge industries.
Proprietary data = proprietary advantage. Why should these firms invest in this manner, when Harvey exists and the frontier labs improve weekly? Because vanilla AI, by definition, regresses to the mean. Law firms that hire the best people and own their case law and playbooks aren't about to feed that edge into a model that hands it to everyone else. Instead, they're using AI to sharpen their secret sauce, not dilute it. Expect this to become the corporate default: generic AI yields average results; AI trained on your moat widens it.
A two-tier legal system? If this works, the white-shoe firms will operate on a different plane from everyone else, from smaller firms, clients, even regulators and the government. The business and societal implications deserve a close watch.
Of course, none of this will stay contained to law. A business model of expensive, highly trained people billing by the hour and sitting on proprietary data also describes consulting, tech services, marketing agencies, and accounting. What's notable is that it has started at scale with law, and it merits a close watch as it's a template the rest of the knowledge economy will copy.
The week's most important AI news, curated for you:
