The Scoreboard Goes Dark: An Epilogue to the DOGE Trilogy on Law Firm Profit Maximization

“The scoreboard that once advertised the strength of the platform now advertises the size of the beast.”

I. The Scoreboard Goes Dark

In the first week of October 2026, the world’s highest-grossing law firm announced that it will stop keeping score in public. Kirkland & Ellis informed The American Lawyer by letter that 2026 would be the last year it supplies revenue and profit figures to the press. Public reporting, the firm wrote, “does not provide meaningful value to our clients,” and the rankings it feeds “could incentivize quantitative metrics over qualitative strengths.” Justin Henry, Kirkland Goes Dark on Reporting Revenue and Profits to Media, Bloomberg L. (Oct. 7, 2026).

The numbers Kirkland will no longer volunteer are not modest. In 2025, the firm reported revenue of $10.6 billion, the first law firm to cross $10 billion, and average profits per equity partner of about $11.1 million. Rachel Scepanovic, Kirkland Pushes Elite Law Firm Profits to New Level as PEP Tops $11m, Nonbillable (Mar. 19, 2026). Within days, Bloomberg Law reported what the letter left unsaid. Firm leaders were wary of how clients would read those figures. Rob Chesnut, Airbnb’s former general counsel, gave the client’s view: “How can you send me that annual letter telling me that expenses are going up and you need to raise your billing rate 10% when I read how much profit the firm is making?” Justin Henry & Meghan Tribe, Kirkland Hides Profits as Big Law Clients Scrutinize Legal Costs, Bloomberg L. (Oct. 9, 2026).

Here is the curious thing. None of this is news. Billing rates have outrun inflation for two decades; Thomson Reuters reports that worked rates have risen 122 percent since 2006. Thomson Reuters Inst., Law Firm Rates Report 2026 (Oct. 2025). Profits per partner climbed right alongside them. Clients complained the entire time, and the trend lines never bent. Nor is Kirkland the first major firm to keep its numbers private. Slaughter and May has never published its results, and Freshfields stopped in 2023. Emily Hinkley, Freshfields Confirms It Will No Longer Publish Its Financial Results, Legal Cheek (July 26, 2023). The only new fact is the decision itself. That fact raises a question worth more than the numbers: why now?

A few weeks earlier, The Economist offered one kind of answer. Big Law, it reported, is being reshaped by partner poaching rather than by artificial intelligence. Between January and mid-September 2026, more than 3,200 partners joined one of America’s 200 top-grossing firms, more than in all of 2021. William Savitt left Wachtell for Gibson Dunn. Cravath, the firm whose name the industry’s organizing model still carries, poached Michael Aiello from Weil Gotshal. Rainmakers switching firms now command $20 million or even $30 million. “You used to hire firms,” a New York partner told the magazine. “Now you hire individuals.” Big Law’s Century-Old Business Model is Being Transformed, The Economist (Sept. 28, 2026).

The Economist is right on the facts. It is incomplete on the cause. The lateral wave and the dark scoreboard are symptoms of the same change, and the change is older and deeper than either one. Something has happened to the way legal work is produced, and therefore to what a law firm is for.

This article is an epilogue. Between March 2025 and February 2026, I published a trilogy in which a fictional young data analyst from “DOGE” visits a fictional 170-lawyer firm. What if DOGE Paid a Visit to YOUR Law Firm? built the economics of hourly cost and profit rates operating at the heart of every hourly-billing law firm. The DOGE Analyst Returns explained the rise of the non-equity partner and its unique contribution to firm profitability. Banishing the DOGE Analyst imagined a venture that walked out the door with several of the firm’s partners. The trilogy told the story of one firm. This epilogue asks what those same forces mean for the industry, using a framework built for exactly this question: the late Clayton Christensen’s theory of disruptive innovation. See Clayton M. Christensen, The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail (1997).

II. Greatly Exaggerated

Big Law’s strengths are real. A large firm absorbs risk. It can put fifty lawyers on a matter by Monday morning. It trains young lawyers systematically; in a way no small firm can replicate. It manages conflicts, malpractice exposure and regulatory overlays at scale. It also performs a job that rarely appears in a pitch deck. Christensen’s later work develops the question of what “job” a customer hires a product to do. See Clayton M. Christensen et al., Competing Against Luck(2016). For many general counsel, one of those jobs is insurance against blame. When a bet-the-company case goes badly, nobody in the boardroom asks why the general counsel hired the firm with the most famous name. That protection has value, and the brand earns its premium in providing it.

Yet those strengths have not deterred predictions of Big Law’s demise. In 2008, Richard Susskind asked whether we had reached the end of lawyers. Richard Susskind, The End of Lawyers? Rethinking the Nature of Legal Services (2008). In 2010, Larry Ribstein pronounced the death of Big Law, arguing that its business model could not survive without fundamental restructuring. Larry E. Ribstein, The Death of Big Law, 2010 Wis. L. Rev. 749. Three years later, Steven Harper described a profession in crisis. Steven J. Harper, The Lawyer Bubble: A Profession in Crisis (2013). The decade that followed delivered record revenue and record profits to the largest firms, year after year. A managing partner who has read a dozen obituaries for the business model is entitled to some skepticism about the thirteenth.

The one vision the prophets shared was that technology would precipitate the downfall of the Big Law firm model. The early predictions focused on what the DOGE analyst called the “IT revolution.” Information technology had a profound impact on the practice of law. Those who practiced as recently as the 1990s had front row seats to the IT revolution. Traditionally, research meant books and bound reporters. Document productions arrived in boxes and were reviewed by teams in war rooms. Briefs and contracts were assembled by hand from form books and internal precedent files. The work required large numbers of people, and the firm that could assemble and coordinate the most people won the largest matters.

The early prophets largely misinterpreted the impact of the IT revolution that followed (online research, email, document management, electronic discovery). The technology did not dismantle the law firm. It made law firms more powerful. Each new tool enabled a large firm to coordinate larger teams across more offices and more data. Electronic discovery, in particular, produced document volumes that only a large firm could review. Technology increased the size and complexity of matters faster than it reduced the labor each matter required. In Christensen’s vocabulary, information technology was a sustaining innovation for Big Law. It improved the incumbent’s performance on exactly the dimensions its best clients valued, and the incumbents embraced it.

The question now on everyone’s mind is how will the AI revolution further shape Big Law? The answer lies both in the technology and in the economic principles that govern law firm profitability, which takes us back to the DOGE analyst’s whiteboard.

III. “I Need to Consult with my Data Processing Center”

Strip away the marble lobbies and the mahogany, and law firms have always been at their core data processing centers. Raw information goes in: case facts, documents, statutes, judicial decisions, a client’s objectives. Processed information comes out: advice, briefs, contracts, arguments. For most of the profession’s history, humans did the data processing. That is why the work required large teams, and why the teams had to sit inside a single large firm.

The first two DOGE articles were an attempt to measure what that processing actually costs and to describe how economic principles shape a law firm’s profit maximization behavior in the face of improving technology. One of the analyst’s key insights was that per-attorney hours, not hourly rates or total revenue or even total hours, fundamentally drive law firm profitability.

He expressed it through two accounting metrics. The hourly cost rate (HCR) is each lawyer’s share of fixed cost divided by the hours that lawyer bills: the hourly “rent” a lawyer pays before profit begins. The hourly profit rate (HPR) is whatever is left of the hourly billing rate after the rent is paid.

The arithmetic is unforgiving because it is non-linear. Fixed costs do not shrink when hours do. Every hour that disappears raises the rent on the hours that remain. But every hour added has the potential to flex profits in the other direction.

This is why “when per-attorney hours are maximized,” the DOGE analyst explained, “HCR is at its lowest and HPR is at its highest, placing firm profits at their zenith relative to the firm’s cost structure.  Likewise, when per-attorney hours are mediocre, so are profits.” What if DOGE Paid a Visit to YOUR Law Firm?, supra.

And that is why, over the decades, technology has presented law firms with both challenges and opportunities in maximizing their bottom lines. Technology has incrementally increased per-attorney efficiencies. Attorneys now do more with less time.

Not surprisingly, per-attorney hours have drifted steadily downward. In 1998, the most common associate billable requirement was above 1,800 hours. NALP, Number of Associate Hours Worked Declines (Feb. 2011). By 2022, Thomson Reuters reported lawyers working about 119 hours a month, roughly 1,430 a year and the lowest level in two decades. Thomson Reuters Inst., 2023 Report on the State of the Legal Market 10 (Jan. 2023). Many firms have struggled to remain profitable. Aside from bringing in more client work, law firm managers have two administrative levers to pull when responding to downwardly drifting per-attorney hours: raising hourly rates and restructuring headcounts.

The first DOGE article pulled the headcount lever while leaving hourly rates unchanged. The first move was a ruthless reduction of associate headcount to achieve optimal per-attorney hours. Profits grew. The next move was “right-mixing” new recruits as the firm returned to its original size. Technology dictated a preference for more senior recruits. The firm’s published hourly rates remained constant, but the average billed hourly rates rose. Profits swelled.

The second article extended the DOGE analyst’s right-mixing strategy by introducing non-equity partners to the firm. Profits ballooned, and the partnership’s superstars sent recruiters to voicemail. The second article also showcased why technology gains largely explain the explosion in annual hourly rate increases across the legal industry.

The article reduced it to an equation: the hourly rate multiplied by project time equals project value. Rearranging the equation, we see the hourly rate equals project value divided by project time. So, where project value to the client stays the same as technology gains shrink project time, the hourly rate must rise. Otherwise the firm fails to capture the project value.

The real market has followed the equation with remarkable fidelity. Top partner rates at large U.S. firms passed $3,400 an hour in early 2026, according to Persuit data, and Susman Godfrey set its top rate at $4,000. Erin Mulvaney, Top Lawyers’ Fees Have Skyrocketed. Be Prepared to Pay $3,400 an Hour, Wall St. J. (Feb. 18, 2026); David Lat, Judicial Notice (02.01.26): The $4,000-An-Hour Club, Original Jurisdiction (Feb. 2, 2026). Rates have risen, in large part, because each matter requires fewer hours, not because clients are buying more of them.

Yet the equation tells only part of the story. If shrinking project time were the whole explanation, rates would rise only enough to hold project values constant. They have risen much further. The reason, as the second DOGE article explained, is that technology has never improved only attorney efficiency. Even the IT revolution significantly improved attorney proficiency. A lawyer armed with searchable databases and electronic document sets did not merely work faster. That lawyer found the case, the admission or the prior art a predecessor never would have found. Efficiency shrinks the time. Proficiency raises the stakes. And when proficiency rises on both sides of a dispute or a deal, the market for legal services stops behaving like an ordinary market.

That is the arms-race logic of the second article, and it explains why clients kept paying. Legal services follow the economics of an arms race, not of an ordinary consumer market. In a consumer market, better technology lowers prices. In an arms race, better technology raises the stakes of losing, and the side with the sharper capability sets the pace which the other side must match. Thomas Schelling and Glenn Snyder taught a generation of strategists that logic; general counsel facing a well-armed adversary learn it the hard way. See generally Thomas C. Schelling, The Strategy of Conflict(1960); Glenn H. Snyder, Deterrence and Defense (1961). A client facing dire consequences hires the best counsel available, and the price of “best” moves with capability, not with cost. Rates therefore capture a sophistication premiumthat better tools create.

The second article also showed how firms capture that premium internally. Technology lets senior associates do work that once required a partner. Each such hour, billed at an associate’s rate, cannibalized a large sum of equity partner profit in the fictional firm. Promoting those associates to non-equity partner status let the firm bill their hours closer to partner rates without diluting the equity. The non-equity partner tier that The Economist attributes to “Kirklandisation” is not merely a compensation fashion. It is a rational response to a changing production function.

Here is where artificial intelligence parts company with the sustaining IT technology that came before it. Information technology expanded what a large firm could coordinate and do. It made a bigger team more productive. Artificial intelligence compresses the processing layer itself: the first drafts, the document review, the research memoranda, the diligence checklists, the very work the standing army existed to perform. It makes a smaller team sufficient. Both IT and AI are efficiency driving technologies. But their effects on Big Law’s profit formula run in opposite directions.

Once the processing layer no longer requires an army, the cost of housing an army changes character. Office towers, deep associate classes, practice groups that exist to fill out a full-service platform: these were wise investments when the work required them. When the work no longer does, they become a tax. Except for the few at the very top, Big Law revenues do not only feed the partners’ profits. They must also feed the beast.

That is the answer to the question of timing. The dark scoreboard has three causes, and they arrived together. First, the absolute numbers have grown large enough that the published profit figure collides head-on with the annual rate letter, which is exactly Chesnut’s complaint.

Second, clients now hold a concrete expectation that artificial intelligence should restrain costs, not merely justify higher rates. A firm that announces record profits, another rate increase and a $500 million AI commitment in the same year has a difficult story to tell.

Third, the scale model has matured. Published profits once served as a recruiting poster in the talent war. With multi-tier partnerships, a single average profit figure now misleads more than it informs; Kirkland’s leaders reportedly worried that it gave clients a false impression of what their lawyers earn. Henry, supra. The scoreboard that once advertised the strength of the platform now advertises the size of the beast.

Withholding the numbers does not change the economics. It removes one visible reminder of them. Firms stop publishing the score when the game itself is changing, and the first sign of the new game is a challenger Big Law already knows well, because it is made of Big Law’s own partners.

IV. The First Challenger: The Split-Off Boutique

The threat lives within each Big Law firm. Take a full-service firm of a thousand lawyers or more, with a dozen practice groups earning very different margins. Within it sits a high-margin group: public company M&A, perhaps, or private equity, or a specialty litigation practice. Its equity partners generate hourly profit rates well above the firm’s average. Yet the firm’s overhead is spread across every practice, and its profits are pooled before they are divided. The high-margin group pays for office space it does not need, for practice groups it never uses and for a brand that is tarnishing in public cost perception. Apart from some marketing and the occasional referral, the group does not need the rest of the firm. The rest of the firm very much needs the group.

That has always been true of some groups in some firms, and the response has always been the same: pay the superstars more and then hope for the best. What has changed is the cost of leaving. In the first DOGE article, the fictional firm’s overhead fell from roughly 30 percent of revenue to roughly 20 percent once it cut excess capacity, and the analyst targeted 18 percent by giving up the office altogether. The partners sublet their building and went virtual. The third DOGE article called the resulting remote, multi-country operation the firm’s “quiet superpower.” A partner who leaves a full-service firm today can start where the fictional firm finished: virtual from the first day, with overhead sized to the work rather than to the platform.

“Any lawyer with a laptop,” the DOGE analyst quipped, “can be a fully functioning law firm.” What if DOGE Paid a Visit to YOUR Law Firm?, supra.

But a partner does not leave alone. The strongest version of the challenger is not a partner who departs with a book of business. It is a team: the equity partners, the non-equity partners and the associates who already work together every day, who share playbooks and trust and the valuable client relationships. The Economist’s headline examples hint at this. Savitt did not move to Gibson Dunn alone; five partners went with him. Five partners followed Aiello as well. Gibson Dunn Recruits Wachtell Co-Chair Savitt Along with Five Other Disputes Partners, Global Legal Post (July 29, 2026); Weil Gotshal Corporate Chair Mike Aiello to Join Cravath, Bloomberg L. (Sept. 11, 2026). Those teams moved from one large platform to another. The split-off boutique simply asks why the team needs a large platform at all.

Once outside, the boutique stays small by design and grows by assembly. When a matter requires technical depth, a language, a jurisdiction or cultural fluency the core team lacks, the boutique partners with a solo practitioner or another boutique for that matter. When a collaboration works, it can become a hire, after what amounts to an extended job interview conducted on real work. The firm expands and contracts at the matter level rather than fixed headcount targets. Its core capability stays constant; its reach changes with each engagement.

The pattern has a historical parallel worth borrowing. The arms races of the last century ran toward bigger and more complex systems: larger fleets, heavier bombers, more elaborate platforms. They ran that way until the cost of matching scale exceeded the advantage scale delivered. Recent conflicts have redirected the race toward smaller, cheaper, more agile systems that can be assembled and reassembled for a mission. Think drone swarms. For a century, the Cravath model favored the larger platform: more lawyers, more offices, more infrastructure. As software takes over more of the legal data processing, advantage shifts toward teams assembled for a particular dispute or transaction. The less friction in assembling them, the better.

The split-off boutique enjoys an advantage that rarely gets the attention it deserves: a smaller conflict footprint. A full-service firm accumulates conflicts across every practice group and every office. An attractive matter for one group is routinely blocked by an unrelated engagement elsewhere in the firm, or by the possibility of offending an important client. In patent litigation, where competitors in an industry sue one another in rotation, those conflicts can close off entire segments of the market. A boutique carries only the conflicts of its own team and its own matters. Clients the parent firm could never represent become available. The boutique does not only compete for the same work at lower cost; it can reach work its parent’s scale has closed off.

The boutique can also offer what clients have requested for decades: predictable fees. In transactional work, a well-scoped flat fee lets the boutique keep the efficiency gains from AI-assisted drafting and diligence as profit, rather than surrendering them as lost hours. In litigation, the most workable structure I have seen divides the case into phases (pleadings, fact discovery, expert discovery, summary judgment and claim construction, pretrial and trial) and sets a flat monthly spend rate for each phase rather than a single fee for the case. Each phase has a bounded scope. The client knows its monthly cost; the firm prices the senior effort it expects plus the technology that leverages it. Big Law can offer the same structures, and some firms do. But its hourly cost rate makes the same fee less profitable or forces leaner staffing than its internal processes prefer. The boutique’s advantage is not the fee form. It is the cost base that makes the fee form sustainable.

Big Law is also, unintentionally, staffing the boutique. The non-equity partner strategy that protects equity profits has created a thick band of lawyers who function at or near partner level and who can see, more clearly every year, how much of their value feeds the beast. Those lawyers are the day-one team of any split-off group. The market already shows that partner-heavy staffing can work at scale: Pierson Ferdinand, launched in 2024 with no associates and with AI tools doing much of the work juniors once did, reported 300 partners by September 2026. Zero-Associates Firm Pierson Ferdinand Hits 300 Partners, Artificial Lawyer (Sept. 30, 2026).

None of this makes leaving easy. A departing group must replace the business development machinery the firm supplied, obtain malpractice coverage on reasonable terms, plan for surges a small team cannot staff and honor the fiduciary and ethical rules that govern departures and client choice. Not every practice combines high margins, portable relationships and low dependence on the rest of the firm. But where those conditions exist, the incentives are real, and technology has lowered the cost of acting on them.

That is what makes the split-off boutique the most dangerous challenger Big Law faces in the near term. It does not merely compete for new work. It removes profitable work and the lawyers who control it. Each successful departure shows the other high-margin groups inside the firm that the dilution they accept is optional. The split-off boutique is both an alternative to Big Law and a mechanism for dismantling it.

And yet, measured against the theory that gives this article its framework, it is not a disruptor at all.

V. Not a Christensen Disruptor (and Why That Is Not Comforting)

Christensen’s theory, as he first set it out, draws a sharp line between two kinds of innovation. Clayton M. Christensen, The Innovator’s Dilemma (1997). A sustaining innovation improves a product on the dimensions the incumbent’s best customers already value. A disruptive innovation does the opposite. It arrives cheaper, simpler or more convenient, and worse on the measures that matter to mainstream customers. It takes root either at the low end, among customers who are overserved by what the incumbent sells, or in a new market, among people who were not buying at all.

The incumbent ignores it, and rationally so: its best customers do not want the new thing, and the margins look unattractive. Then the new thing improves, moves upmarket and one day it is good enough for the customers the incumbent thought were safe. Christensen also found that incumbents almost always win the sustaining battles. They have the resources, the customers and the motivation.

Measured against that definition, the split-off boutique is not a disruptor. It is the opposite. It targets the incumbent’s best clients, with the incumbent’s best lawyers, at the incumbent’s level of quality. It is a sustaining entrant that happens to have a lower cost structure. Christensen’s prediction for such entrants is plain: the incumbent fights and usually wins, or it absorbs the challenger.

The framework has a better tool for describing what the boutique actually is. In The Innovator’s Solution, Christensen and Michael Raynor explained when integrated businesses win and when modular ones do. Clayton M. Christensen & Michael E. Raynor, The Innovator’s Solution ch. 5 (2003). While a product is not yet good enough for what customers need, the integrated company wins, because it controls every interdependent piece. Once performance overshoots what customers can use, the basis of competition shifts to speed, flexibility and price, and modular architectures win. The full-service firm is an integrated architecture. The swarm assembled for a single matter is a modular one. “We hire lawyers, not law firms” is a client describing a modular interface. Artificial intelligence pushes more of the processing layer past the point of “good enough,” which is precisely the condition under which modularity takes over.

The same book supplies a corollary that explains the money. Christensen called it the law of conservation of attractive profits. Id. ch. 6. When one layer of an industry becomes modular and commoditized, profits do not disappear. They migrate to an adjacent layer where performance is still not good enough. That is a fair description of what The Economistreported without naming. Pay packages of $20 million and $30 million are profits migrating away from the institution and toward the layer that remains scarce: elite judgment and the client relationships that travel with it. Profits will migrate in a second direction as well, toward whoever owns the integration layer that technology is now building. That layer is where the most serious threat will form.

Christensen’s prediction about sustaining entrants can be tested, and those of us in patent litigation lived through the test. In the 1990s, specialized intellectual property firms dominated the field. Over the next two decades, many of them were absorbed or disappeared. Pennie & Edmonds dissolved at the end of 2003. Fish & Neave merged into Ropes & Gray in 2005. Morgan & Finnegan’s lawyers went to Locke Lord in 2009. Darby & Darby dissolved in 2010. Townsend and Townsend and Crew joined Kilpatrick Stockton in 2011, Kenyon & Kenyon’s lawyers joined Andrews Kurth in 2016, Fitzpatrick, Cella, Harper & Scinto joined Venable in 2018, and Brinks Gilson & Lione joined Crowell & Moring in 2021. Christensen would have nodded.

But not every boutique followed the script. Finnegan, Fish & Richardson, Knobbe Martens and Sterne Kessler remained independent and grew. Quinn Emanuel offers a second counterexample at the top of the market. Founded in Los Angeles in 1986 as a litigation-only firm, it never became full-service. It now reports revenue of roughly $2.8 billion and ranked fourth in the 2026 Am Law 100 for profits per equity partner. Quinn Emanuel Posts Third Year of Double-Digit Growth as Revenue Hits $2.8bn, Legal Business (Mar. 25, 2026); The 2026 Am Law 100 Is Out, Above the Law (Apr. 2026). A specialist that declined to integrate did not merely survive the integrated giants. It joined them in the top tier.

Those histories are instructive for a second reason. Most of the absorptions happened during the IT revolution, when growth required what only a platform could provide: offices, litigation support, armies of reviewers for electronic discovery, and the capacity to staff the largest matters. Boutiques were pulled into platforms because scale was the price of growth. Artificial intelligence weakens that pull. If a forty-lawyer firm can staff a case that once required a two-hundred-lawyer firm, the gravitational argument for joining a large platform loses much of its force.

Here is why the boutique’s failure to qualify as a Christensen disruptor offers Big Law little comfort. Christensen’s corporate incumbents won sustaining battles because they owned the resources needed to win them: the plants, the patents, the engineers’ work product, the customer contracts. A law firm owns almost none of its essential resources. Its lawyers are free to leave; the profession’s ethics rules forbid most restrictions on their right to practice elsewhere. Its client relationships belong, in practice, to the lawyers who hold them. The resources go home every night, and in a virtual firm they never come into the office. A disk drive maker could choose to win a sustaining fight. A partnership cannot win a sustaining fight with a team it cannot keep.

The split-off boutique, then, is the first serious challenger Big Law will face, but it is a challenger inside Big Law’s own game. True disruption, if it comes, will not look like a law firm at all.

VI. The True Disruptors

Christensen insisted on a point that is often lost in popular accounts of his work: disruption is a business model, not a technology. The same technology can sustain one company and disrupt another, depending on which company’s profit formula it fits. Artificial intelligence is not disruptive in itself. What it does is make viable a set of profit formulas that the Big Law pyramid cannot easily exploit without dismantling itself. A firm that earns its profit by the hour cannot sell outcomes, subscriptions or fixed prices at scale without cannibalizing the hour.

The legal market has already lived through one complete Christensen disruption, and Big Law behaved exactly as the theory predicts. Electronic discovery was, at first, a sustaining bonanza: review teams billing by the hour, by the hundreds, on matters of unprecedented volume. Then contract reviewers, outsourcing providers and technology-assisted review took the work away from the low end, priced per document or per gigabyte. Big Law let it go. The margins were thin, and the firm’s best clients had no interest in paying partner-track rates for first-level review. When Christensen’s 2013 article said disruption had “recently come for law,” this is the kind of migration it described. Clayton M. Christensen, Dina Wang & Derek van Bever, Consulting on the Cusp of Disruption, Harv. Bus. Rev., Oct. 2013. The question is not whether disruption can happen in law. It already has. The question is what climbs the value chain next.

Christensen’s work suggests five questions that separate a true disruptor from a well-funded competitor. Does it serve customers who are overserved, or who are not buying at all? Does it make money in a different way? Is its product good enough for some uses, and improving quickly? Would the incumbent’s best customers reject it, at least at first? And does the incumbent, looking at it, see something unattractive rather than threatening? Applied to the challengers now visible in the legal market, the questions separate the merely competitive from the genuinely disruptive.

The first true disruptors will serve people who are not buying legal services today. Consider the mid-market company that holds valuable patents but cannot fund a multi-million-dollar infringement case. Or consider the Japanese mid-size manufacturer that owns U.S. patents and has never once considered enforcing them, because the cost, the distance and the unfamiliarity make enforcement unthinkable. An AI-enabled shop that makes enforcement viable for those owners, at a fixed or outcome-based price, takes nothing from Big Law at first. Those clients were never Big Law’s. That is precisely why Big Law will not notice.

Capital is the second ingredient. Patent enforcement already contains the embryo of a disruptive model: litigation funding paired with lean trial teams. Add artificial intelligence to the team and the funder’s economics improve with every hour the software saves. Capital also dissolves a constraint that binds every partnership: the need to pay for long-term investment out of this year’s distributions. Funders understand this. Burford Capital has said publicly that it is pursuing minority stakes in U.S. law firms through alternative business structures and management services organizations. Burford Aims to Buy Stakes in US Law Firms, Bloomberg L. (Aug. 18, 2025).

The AI-native firms attack from the low end. Crosby, which raised a $60 million Series B in 2026, sells contract review. General Legal serves startups. Eudia Counsel obtained an Arizona ABS license in 2025 for contracting and diligence work. The Economist noted David Fox, formerly of Kirkland, who claims his “AI-first” firm’s software performs as well as a lawyer on the verge of partnership. Most of these firms sell work that Big Law regards as beneath its rates. One, Norm Law, launched in late 2025 by the compliance-AI company Norm Ai and chaired by Mike Schmidtberger, who led Sidley’s executive committee from 2018 to 2025, is already testing whether the model can move upmarket into financial-services regulatory work.

Two further forces complete the picture. The first is the client. In-house legal departments armed with the same tools are taking over the coordinating role that justified the full-service firm, buying specialists directly and assembling their own swarms. That disrupts Big Law’s integration function rather than its lawyers, and integration is where the conservation of attractive profits says the money will flow. The second is the Big Four. KPMG Law US received an Arizona ABS license in February 2025, the first Big Four firm licensed to practice law in the United States, and it began with high-volume contracting and managed legal services: textbook low-end entry, backed by capital no law partnership can match. Bob Ambrogi, KPMG Becomes First of Big Four to Practice Law in U.S. as Arizona Approves Its ABS License, Legal.io (Feb. 2025).

Then there is the disruptor that does not yet exist. Christensen’s third principle of disruption is that “markets that don’t exist can’t be analyzed.” Christensen, The Innovator’s Dilemma, supra. The most consequential disruptor may be one nobody can describe today. Disruptors need not be young, but they must be unburdened by the assumptions of the pyramid: that legal work is priced by the hour, produced through leverage and owned by partners. In Banishing the DOGE Analyst, the venture that upended the fictional firm was conceived by a young man who had never practiced law and saw no reason the work had to be organized the way it was. Big Law’s blind spot is not technology. It is its own definition of “real legal work.”

Disruption, if it arrives, will be uneven. Process-dense work (contracting, diligence, routine discovery, fund formation) is exposed first. Bet-the-company litigation, multi-jurisdictional crises and matters that demand deep institutional risk-bearing will be exposed last, if at all. Regulation will shape the geography. Model Rule 5.4 still bars non-lawyer ownership in nearly every state. Arizona abolished its version of the rule effective 2021 and has approved more than 150 alternative business structures. Utah narrowed its regulatory sandbox in 2024 and has set it to expire in 2027. On October 2, 2026, the Tennessee Supreme Court declined to permit non-lawyer ownership but said it may revisit the question as data accumulates. Order, No. ADM2025-01403 (Tenn. Sep. 16, 2025). The proposal is under active consideration in more than one state. In some, it may be only a vote away. The signals worth watching are group rather than individual laterals, the formation rate of boutiques founded by Big Law partners, pressure on realization at premium rates, and requests for proposals that trade brand for team continuity and total cost.

VII. “The Platform”: An Instructional Fiction

The DOGE trilogy closed with a fictional venture called “the Platform.” It remains fictional. Its value here is instructional, because almost none of its parts are themselves fictional.

In Banishing the DOGE Analyst, the analyst returns to the firm’s partner retreat with a product demonstration. The Platform is an integrated system with persistent matter memory and continuous monitoring. It cross-references deposition testimony against the record in real time, flags appellate decisions the moment they issue, builds risk matrices for merger diligence and trains junior lawyers through a module modeled on the reasoning of senior practitioners. Each capability, the analyst insists, uses technology that exists today. What the Platform adds is integration. A premium module scans public data and identifies disputes and transactions before the affected companies have retained counsel: a client development bonanza.

The business model is even more threatening than the product itself. The Platform begins as a vendor, selling subscriptions priced like a research service, with practice-specific add-ons and usage charges passed through to clients. A vendor, however, lives at the mercy of its customers’ budgets. So the Platform escapes. It first offers to carry the cost of its services on a litigation in exchange for a stake in the outcome and then begins investing in cases directly. It forms a law firm subsidiary as an Arizona alternative business structure, able to raise outside equity and to grant ownership stakes to engineers. Its path into premium work runs through a merger with a respectable mid-sized firm. Along the way it discovers a strategic weapon in the conflict rules: the first party to engage the Platform on a matter, as the analyst puts it, handcuffs the other side.

Read it again as a list of components, and every one of them exists. Arizona’s licensed alternative business structures already include KPMG Law US and Eudia Counsel. Harvey, which says about four in five Am Law 100 firms use its tools, raised capital at a $15.5 billion valuation in September 2026, and its arrangement with A&O Shearman to build and sell agentic tools gives a law firm a share of software revenue. Harvey, Harvey Raises $550M at a $15.5B Valuation (Sept. 9, 2026). Outside capital is already flowing into legal services: Norm Ai has raised more than $260 million from investors that include Blackstone and has launched its own law firm. Litigation funders already take stakes in outcomes, and Burford Capital is seeking stakes in law firms themselves. Nobody has yet assembled these pieces into a single enterprise. The fiction lies only in the assembly.

Run the Platform through the five questions, and it passes each one. It creates a market, finding and funding disputes that would otherwise go unbrought. It makes money from subscriptions and, decisively, from outcomes rather than hours. It is good enough for some uses today and improving with every matter it absorbs. The largest clients will not trust it with a bet-the-company case at first. And Big Law will see it as a vendor rather than a rival, which is the most dangerous perception of all. Firms will license the tools, and in licensing them they will help fund the enterprise that competes with them.

The move from subscription to stake is the decisive step. A vendor’s profit is capped by what its customers will pay, and its customers are law firms whose own resource allocation keeps technology budgets lean. A principal’s profit scales with the outcomes it finances. Once the Platform holds stakes in the matters it serves, law firms stop being its customers. At best, they become its suppliers. That is the law of conservation of attractive profits operating in plain view: the layer that integrates captures the profit, and the layers it integrates compete on price.

The Platform also has a ceiling, and the ceiling is instructive too. Exclusivity cuts both ways. A Platform that handcuffs one side of a dispute cannot serve the other. A single Platform therefore cannot dominate an entire market; the conflict rules all but guarantee competitors. Regulation adds friction. Alternative business structures remain confined to a few jurisdictions, and the fictional firm discovered that its multi-state, multi-country practice did not fit neatly into Arizona’s framework. The likely end state is not one Platform but several: a market structure Big Law can join or supply, but not control.

Now combine the Platform with the challenger from Section IV. Separately, each has a weakness. The split-off boutique lacks the technology investment of a large platform. The Platform lacks elite lawyers and the client relationships that travel with them. A split-off team that rents the Platform has integration on the scale of a large firm without the cost of feeding the beast. Drone swarms defeat superpowers only when they are linked by command and control. The Platform is that network.

Nothing in this account is far-fetched. A venture built this way might be enormously profitable for its lawyers and its outside investors. And almost none of it has anything to do with the Cravath model. The Platform does not beat the Cravath model at its own game. It ignores the game. That is the instruction the fiction offers: the most serious threat to Big Law may come from an enterprise to which leverage, the billable hour and the equity partnership are simply irrelevant.

VIII. Big Law Fights Back

Big Law is not standing still, and a fair account must give its responses their due. Two deserve close attention: the continued up-mixing of the lawyer pyramid, and the large investments some firms are now making in proprietary artificial intelligence.

Up-mixing is the strategy the second DOGE article modeled: fewer junior associates, a thicker band of non-equity partners and senior associates doing work once reserved for partners, and equity concentrated among the lawyers who control clients and high-end judgment. It works. It raises blended rates, protects realization on sophisticated work and holds profits per equity partner steady or higher even as hours per matter fall. It is also a textbook sustaining innovation, and Christensen’s central finding is that sustained success at sustaining innovation predicts later vulnerability, not immunity.

Every turn of the up-mixing screw elevates more lawyers to near-partner capability and near-partner expectations. Every turn makes the economics of leaving easier for those lawyers to see. The strategy that defends the platform also recruits the boutique.

The proprietary AI investments are more interesting, and a skeptical managing partner should take them seriously. Kirkland has reportedly committed $500 million over three to four years to build its own AI platform. Eric Killelea & Roy Strom, Kirkland & Ellis Investing $500 Million to Build AI Platform, Bloomberg L. (May 28, 2026). There are two good reasons to think this is smart.

The first comes from Christensen himself. Integration wins while a product is not yet good enough, and legal AI is not yet good enough for the highest-stakes work. A firm that builds a tightly integrated system around its own work product, rather than licensing a generic tool, is making the correct move for this stage of the technology. The second reason is subtler. Proprietary AI is the first firm-owned asset of any size that does not go home at night. A system that holds years of matter memory, templates and judgment-tested work product could make a lawyer more effective inside the firm than outside it. If it does, the institution is once again appealing, and the client who hires the lawyer must also hire the firm. That is a direct answer to the split-off boutique. A departing group can take its clients. It cannot take the system.

The limits are equally real, and the DOGE economics expose them. Start with cannibalization. Hourly rate times project time equals project value. Inside an hourly model, a tool that reduces hours reduces revenue unless the rate rises to compensate. The firm’s resource allocation process will therefore steer the tool toward work that justifies higher rates rather than work that lowers the client’s total cost. The tool will be built to sustain, because that is what the partnership can afford to build.

Then consider how the investment is funded. A law partnership pays for technology out of the same year’s revenue that would otherwise be distributed to its partners as profit. Kirkland’s commitment is large in absolute terms. But spread over three or four years it amounts to a little over one percent of annual revenue, a cost center absorbed by a single firm and amortized across a single firm’s matters.

A stand-alone Platform amortizes its technology across many firms, corporate legal departments and other professional users, and treats it as a capital asset financed by investors with a long horizon. Harvey raised $550 million in a single funding round in September 2026. A proprietary system must also stay ahead of commercial models whose development is financed on a scale no law firm can approach. When commercial tools become good enough, Christensen’s modular moment arrives, the premium on proprietary integration erodes and the boutique can rent most of the same capability by the month.

The alternative carries its own warning. In 1981, IBM built the personal computer and outsourced its microprocessor to Intel and its operating system to Microsoft. The suppliers, not IBM, captured the industry’s profits. Christensen & Raynor, supra, ch. 5. A firm that licenses its AI layer rather than owning it risks becoming IBM. A firm that builds its own avoids that fate only if it can keep pace.

Underneath both strategies lies a constraint peculiar to the partnership form. In a corporation, a long-term investment can be financed with retained earnings, new equity or debt, and its payoff is reflected in a share price that every owner can sell. In a law firm partnership, the investment comes out of current distributions, dollar for dollar, and the ownership interest cannot be sold.

Partners therefore experience the same investment very differently depending on where they stand in their careers. An equity partner three years from retirement captures little of the payoff from a system that matures in five. A younger equity partner may doubt still being at the firm, or still in the equity, when the payoff arrives. This dynamic is frequently on display in a decision that has nothing to do with technology: whether a firm should buy or lease its building. The long-lived asset makes sense for the firm as an institution. But the partners who would pay for it have different horizons and different exit options. Proprietary AI is the building decision again, with higher stakes and a faster depreciation schedule.

The menu of realistic responses follows from these constraints. Firms can layer sustaining technology onto the existing model. They can create innovation units that remain tied to the partnership’s economics. They can experiment with alternative business structures or management services organizations where the law allows. They can tolerate, or even facilitate, the departure of a high-margin group in exchange for referral arrangements. And they can tighten their defenses: origination credit rules, multi-year guarantees for stars, more non-equity tiers.

Christensen’s logic predicts which will dominate. The first and the last will, because they fit the partnership’s processes and values. The middle options will be attempted and under-resourced. Facilitated spin-outs will be resisted until groups begin leaving anyway. None of these responses removes the incentive for a high-margin group to leave, unless the firm is willing to shrink its own scope. Few partnerships will choose that voluntarily.

IX. The Innovator’s Solution

Christensen did not leave incumbents without hope. The Innovator’s Solution prescribes a specific remedy: an autonomous organization, with its own resources, its own processes, its own cost structure and its own profit formula, insulated from the core business’s resource allocation. The insulation is the point. Inside the core business, every dollar for the new venture competes against the demands of the best existing customers, and the new venture loses every time.

Translated to a law firm, the textbook answer is uncomfortable. The firm would spin its technology out into a separately capitalized company, or into an alternative business structure where the law permits one. That entity would sell to other firms and to corporate legal departments, take outside investment, and eventually hold stakes in the matters it serves. The firm would, in effect, build its own Platform before someone else builds one around it. The A&O Shearman arrangement with Harvey, in which the firm shares in revenue from tools sold to others, is a modest step in that direction.

The obstacles are formidable. A technology company owned by a law firm would sell to the firm’s competitors and to its clients’ adversaries, raising conflicts questions no ethics committee has fully answered. Outside of Arizona and a handful of other jurisdictions, Rule 5.4 limits who may own what. And every partnership would have to decide who receives equity in the new entity, a question with the power to divide a firm faster than any lateral raid.

The trilogy suggested a partial way through. In Banishing the DOGE Analyst, several partners did not wait for a firm-wide vote. They invested in the analyst’s venture personally. The detail seemed like plot at the time. It turns out to solve a real problem. Personal investment dissolves the career-stage constraint, because each partner chooses according to a personal career horizon instead of the partnership imposing a single horizon on everyone. The retiring partner can decline. The forty-year-old partner can commit. A firm could sponsor such a vehicle, invite voluntary participation from its partners alongside outside investors, and keep the core partnership’s distributions intact. The ethical and governance questions are real, but they are questions of design, not of principle.

One more problem remains, and it is the one the fictional DOGE analyst raised most pointedly: “If you eliminate the training ground, you eventually eliminate the profession.” The split-off boutique, the up-mixed platform and the AI-native firm all depend on lawyers who learned their craft in the pyramid. Every one of them is, in a sense, free-riding on Big Law’s investment in training. If the pyramid thins, someone must still train the next generation of lawyers good enough to be worth assembling into swarms. The Platform’s answer in the trilogy was a training module modeled on the reasoning of senior practitioners. Whoever solves training at scale, whether a firm, a Platform or a law school, may end up owning the most durable asset in the legal market: the next generation of lawyers. That opportunity is open to Big Law. It may be the one place where scale still confers an advantage no swarm can easily match.

X. Lights Out

A scoreboard exists for a particular kind of game: one played on a fixed field, between known teams, under agreed rules. For a century, Big Law played that game. Firms competed on size, on reach and on profits per partner, and the Am Law tables kept the score.

Kirkland’s decision to switch off its scoreboard is easy to dismiss as public relations. I read it as something more significant: an acknowledgment, perhaps not yet fully articulated, that the game itself is changing. The challengers now taking the field do not play for the same score or even by the same rules. A split-off boutique measures success by what its partners keep, not by what a firm reports. An AI-native firm measures it in matters closed per lawyer. A Platform measures it in returns to its investors. None of them needs the scoreboard, and few of them will ever appear on it.

The DOGE trilogy ended on a tarmac, with a fictional managing partner watching several of his partners board a plane as they left the firm and partnership, and with a question for the reader: what would you do? The question is no longer fictional. Every managing partner now faces the choice the fictional firm faced, with real partners, real clients and real capital in motion. The question was never whether this moment would arrive. It is what each firm will do now that it has.