The $500 Million Answer
Monday, June 1, 2026
Last week the highest-grossing law firm in the world told the legal AI market it had decided to build rather than buy. The number is the headline — $500 million over three to four years, with more than $100 million spent this calendar year — but the decision underneath it is the story. Kirkland & Ellis is not unhappy with the AI tools on the market. The firm is buying those too. What it does not believe is that any of them, when used by competitors simultaneously, can produce a strategic edge. So, it is building its own.
Kirkland is not alone in the move, and that is more significant than the dollar figure. The same week, three other elite firms made parallel decisions — at a smaller scale, with different framing, pointed in the same direction. The market has split. A handful of firms at the top are building. Most are buying. The question for everyone in between is no longer whether to invest in AI, but how to invest in a way that creates something competitors do not have.
May 28, 2026: Kirkland's $500 Million Bet
The Financial Times broke the story. Kirkland & Ellis is committing $500 million over the next three to four years to develop a proprietary AI platform from the ground up. Not a Harvey deployment. Not a Copilot rollout. Not a vendor partnership in a co-marketing wrapper. A platform engineered by the firm's own data scientists, with input from 250 of its lawyers, including 100 equity partners, designed to handle complete mandates end to end. The platform will not be licensed, sold, or shared. Outside vendors are involved in the build, but they are prohibited from selling the resulting work to other firms. The investment is being funded from revenue — Kirkland generated $10.6 billion last year — and starts with more than $100 million in 2026.
The framing came from Chairman Jon Ballis, and it is the most clarifying sentence anyone in the legal AI conversation has said all year. Widely available tools like Harvey, Ballis told the FT, are "raising the floor for everyone" in the legal industry. That is not criticism. It is a compliment to the vendors and a precise explanation of why, for a firm that competes at the ceiling, using only those tools would be a strategic mistake. If every Am Law 50 firm has access to the same Harvey instance and the same iManage context fabric, then the AI does not differentiate among them. It just becomes the new baseline of competence. Kirkland's bet is that the firms that will pull ahead are the ones that build something competitors cannot copy.
There is a precedent worth noting. About a decade ago, Kirkland built a proprietary database called CTRAN to track terms across past M&A transactions and used that data advantage to spot trends and price deals more sharply than competitors who relied on industry reports. The $500 million AI commitment is the same instinct at a much larger scale: control the technology that creates leverage, do not rent it.
What Kirkland will not say in public is the harder question the investment opens. AI infrastructure is not a database. It is a continuously evolving stack — models, retrieval, governance, evaluation, fine-tuning, training data, agent frameworks — that needs to be maintained at a velocity no law firm has previously matched. Building the platform is the visible part. Operating it for a decade against a moving frontier is what decides whether the $500 million was a strategic asset or an expensive lesson.
The Same Week: Fried Frank, Linklaters, Littler Mendelson
Kirkland's announcement was the loudest, but it landed inside a cluster of similar moves that reframe the story.
On the same day Kirkland's news broke, Fried Frank announced it was rolling out a new, internally built AI platform to streamline its private equity practice. The firm did not put a number on it, but the structural choice is the same: build the tool within the firm, designed around the firm's actual workflows rather than a vendor's product roadmap.
Earlier in May, Linklaters launched a dedicated team of data scientists and lawyers who will build bespoke AI workflows for specific clients and matters — not a single platform for the whole firm, but a workshop approach where AI is custom-fitted to the work at hand. In April, Littler Mendelson — the largest employer-side labor and employment firm in the US — created a top executive position to steer the firm's internally developed AI tool.
Four firms, four different builds, one direction of travel. None of them is walking away from the vendor market. They are continuing to license Harvey, Microsoft Copilot, and the rest of the stack. What they are saying is that off-the-shelf alone is not enough at their tier of competition, and the right response is to build proprietary capability on top of it.
In Other News This Week
The Illinois State Bar Association announced it has selected SimpleDocs as the AI vendor for its new member benefit program — among the first state bars to offer legal AI as a membership perk. It is the buy side of last week's story, and a useful counterpoint: an institution serving thousands of solo and small-firm members concluded that picking a vendor was exactly the right move for its constituents.
Align Technologies expanded its mobile litigation platform with a new offline read-write mode for iPad, aimed at trial teams working in courtrooms and depositions where connectivity is unreliable.
European legal AI startup Nu:legal raised €1.3 million in seed funding, a smaller round in a week dominated by nine- and ten-figure numbers, and a reminder that the legal AI market is still attracting capital across every layer.
Harvey announced it is opening a Chicago office in July, its second major US location.
The CTO Gap Zone™ of the Week: Build, Buy, or Build a Position
Kirkland's $500 million is the wrong number for the question most firms face. The right number is whatever the firm has to spend on technology over the next three years, and the question is how to deploy it strategically rather than reactively.
For the vast majority of firms — Am Law 100 down to single-shingle practices — the answer is not to build a proprietary AI platform. The capital, the engineering talent, and the operational discipline to run one are out of reach. That is the easy part of the answer. The harder part is that buying everything is not a strategy either. A firm that licenses Harvey, Copilot, an iManage context fabric, a CLM platform, an e-discovery tool, and three vertical AI applications has not made a strategic decision. It has made a procurement list. When every other firm at its tier has the same list, none of them have built anything proprietary.
The defensible middle is to build a position, not necessarily a platform. That means deciding which two or three areas of the practice will be genuinely differentiated by how the firm uses technology, then investing real money, real internal expertise, and real change management into making those areas exceptional. Everything else can be bought at market terms. A litigation boutique might invest heavily in proprietary trial preparation workflows on top of standard e-discovery tools. A mid-sized M&A practice might build its own deal-precedent system on top of a vendor's DMS. A regional firm might invest in a client-facing intake and matter-management experience that competitors cannot match. None of those require $500 million. All of them require a deliberate strategic choice, sustained investment, and the willingness to say no to the temptation of buying everything that arrives.
The Kirkland announcement will put pressure on firms across the market to take visible action on AI. The wrong response to that pressure is more procurement. The right response is a real strategic conversation about where the firm competes and where technology can compound that position.
What This Means for Your Firm
Kirkland set a new ceiling, not a new standard. Most firms will never spend $500 million on anything, let alone software. But every firm now operates in a market where the floor keeps rising — pushed up by the same vendor tools every competitor can buy — and the ceiling keeps moving further out of reach, defined by what the elite tier is now building privately. That is an uncomfortable middle to compete from, and it is where most of the market lives.
It does not mean copying Kirkland. It means recognizing that the era of "AI strategy" being defined as a vendor evaluation is ending. The firms that do well in the next few years will be the ones that treat technology decisions the way they treat lateral hires and practice investments — strategically, deliberately, with a clear theory of where the firm is going and what it will need to get there. The technology budget stops being an IT line item and starts being a strategic capital allocation.
That is the conversation Kirkland just made unavoidable. Firms can no longer wait to build a strategy instead of a tech stack.
The experts at Legal Tech Decoded can help. Reach out to us to schedule a complimentary Strategy Discovery Call.
Laura Kennedy and Chelsey Lambert are Co-CEOs and Co-Founders of Legal Tech Decoded. Rising Above the Noise lands every Monday at legaltechdecoded.com.
