The pricing of legal services in 2026

Before we can determine whether AI is having an impact on pricing, we first have to cover the current state of the market. After all, if one's trying to see whether something is impacting an ecosystem, one must first understand whether that system is already leaning toward feast or famine. In the case of the average large law firm, the feast is very much happening. The average standard rate¹ for US firms is a startling $800 per hour,² with that rate dropping to $720 after initial discounts are applied.

‍

The prices of legal services have been surging for the last few years — a shift, notably, that did not begin with the trumpet-accompanied arrival of ChatGPT onto the world stage in late 2022. Rather, it began with the pandemic itself, which ushered in a period of price volatility that has uncapped rate growth for most of the years since. During those years, law firms of all sizes found they could raise rates two or even three times as much as they once could, first because inflation³ pushed up all prices in 2022-23, then because firms needed to catch up for lost ground. Now, firms are raising rates for different reasons altogether, most likely related to AI and the anticipation around it. That makes this a distinct enough stretch of the market to deserve its own name, and we have given it one in this report: the AI Era.

The segments decouple

It’s important to note though that not every tier of firms has seen the same ability to increase their prices, and 2026 put that in the freshest, starkest perspective. Law firms within the Am Law 100 were able to press massive increases in rates, growing by a new record 9.7% over the previous YTD June levels and notching a fourth consecutive year of rate growth above 7.3%, a feat no other segment has ever managed to achieve. Am Law 100 firms have been leading the charge for the better part of a decade now, pressing rates higher and higher. Each year, people question if this is the year they blink as client pushback finally materializes. Well, 2026 was not that year.

In normal years, that would be the end of it. For the better part of two decades, the segments have been in lockstep, following the same economic beats, accelerating and decelerating together with each year’s rate schedule.

Yet for the first time, someone blinked — and it was the segment that was already having the greatest troubles, midsize firms. The midsize segment decelerated its worked rate⁴ growth, bucking the ever-higher trend of the largest firms. This is not to say that midsize firms lowered their rates (they still grew by 5.4% on average, ahead of inflation), but rather they simply raised them less quickly than the year before.

Perhaps this was intended to work to the advantage of these firms, which have been growing demand far faster than their peers for years now on the back of mobile, price-conscious demand that they’ve been able to siphon off the ever-more expensive top law firms. Or, by backing off on increasing rates so aggressively, midsize firms may be hoping to lure more customers. Either way, the die is cast.

As for the final segment, the Am Law Second Hundred, it decided to neither accelerate nor decelerate, increasing their rates basically at the same pace it did in 2025.

‍

Despite this split, the overall result has been an unprecedented period of growth for law firms, with worked rates growing by 138% since pre-GFC 2006.

‍

Rate increases are a year-to-year phenomenon. And while they track longer-term macroeconomic forces, they can also shift quickly when market conditions do. Underneath them sits a slower force with a subtle but important effect on price: leverage, meaning the title composition of the firm itself. That composition has been evolving for a long time, and that evolution predates the pandemic by years.

For example, large law firms have become increasingly leveraged, both in terms of adding more associates per equity partner, but also by enlarging their non-equity partnership. The logic is straightforward: Every timekeeper who bills more than they cost contributes margin to a partnership that is being divided among proportionally fewer people. This makes leverage a reliable lever on profit per equity partner, a critical metric.

What makes leverage worth a closer look now is that much of the talk around AI may be presaging a total directional change. The argument runs cleanly enough: The tasks that fill a junior associate's day — such as document review, first-draft research, and diligence — are precisely the tasks generative AI (GenAI) performs best. If that holds, firms will need fewer juniors for the same book of work, potentially resulting in a thinner senior cohort supervising machines instead of people. Further, rates should climb as a result, since the mix would then tilt toward high-charging partners and toward the low-volume, high-rate work that cannot be automated away.

However, that is not what we're seeing — at least so far. Associate-to-equity-partner leverage now sits at 1.3 associates per equity partner, up from the 1.13 that held through 2018 and 2019. The non-equity tier moved harder still, to 0.94 as of Q2 2026, from roughly 0.70 in 2018. And that pattern holds whether the ratio is built from headcount or from hours.

Together, that is a rise to 2.24 leveraged lawyers per equity partner, from about 1.83. The two purchases do behave differently on the invoice: associates buy leverage at the cost of dilution, since their rates sit far enough below partner rates that a rising share of their hours pulls the blended rate down; while non-equity partners bill close enough to equity partners to deliver the same leverage without the same drag.

However, firms did not choose between them; rather, they bought both and kept buying. Nothing here is signaling a reversal.

The forecast, made loudly and often since 2023, was that this would dramatically change. If GenAI absorbed the document review, first-draft research, and diligence that fill a junior associate's day, firms would need fewer juniors. Yet, the data says the opposite.

There are more associates per equity partner now than at any point before the pandemic, and firms have added internal tiers between the entry and equity classes rather than removing them. Set that next to the title-level rate figures, in which associate increases trailed non-equity partners by less than one percentage point, and the behavior is consistent in both directions. Firms are not pricing leveraged hours as though they are getting cheaper to produce, and they are not staffing as though they need fewer junior lawyers. Whatever firms say they believe about AI, they are still buying junior labor — and buying more of it than before.

The costs (or lack thereof) of rate increases

A firm can set its rates wherever it likes, and it may even hold the line on a matter or two, but if they aim too high across the board, clients will simply take their work to firms willing to price more modestly. In that way, legal demand is what tells us the real health of the market once the rates have been set — and for law firms in 2026, the news is almost universally good.

The average law firm's demand is up 2.8% year-to-date in 2026 compared to 2025, with performance actually accelerating from Q1 to Q2 despite ever-higher baselines to clear. Whether that surge of work is being driven by regulatory upheaval, evolving technology, or international instability, it has left firms busier than ever, even as their rates kept climbing. Performance is strong enough that productivity — measured as hours worked per lawyer — looks set to end the year flat or slightly up compared to last year, a feat the industry has statistically managed only about one-quarter of the time.

What makes 2026 genuinely notable is that demand is once again doing real work on firms' pre-realization performance, something it had barely done in years. Following declines in 2008-'09, law firm demand went essentially flat, and the decade that followed was less a recovery than a long rebuild toward those old highs, with rate increases doing nearly all the lifting on revenue. The first years of the AI era have not merely held that pace, they've broken it open.

‍

Demand has grown 11% since 2023. Set that against a 2006 baseline that took 17 years to clear, and against worked rates that climbed 138% since that 2006 baseline, and the shape of the change is obvious. Fees worked are now rising on two main drivers — rising rates and increasing demand — instead of one. And that means, price is no longer the only lever firms have.

However, let’s look at what is fueling that second engine. Regulatory upheaval, technological disruption, and international instability are all driving demand. And whatever weighting you assign them, they all represent the same commodity: Risk. Clients are not buying more legal services because they developed a taste for them. Rather, they are buying because exposure went up, and exposure has to be managed whether or not the budget likes it.

That makes this demand unusually sticky. A general counsel facing an enforcement action or a ruptured supply chain cannot decide to forgo these issues until a later quarter.

And that leaves clients in an awkward position. After years of compounding rate increases, and with the volume lever largely bolted in place, the only lever clients may have is the invoice itself, and the option of not paying all of it. Realization is therefore where you would think client pushback would surface first.

‍

So far, however, it isn't surfacing. None of the three major law firm segments show a meaningful decline. Worked-to-collected realization — which measures how much of a matter's value a firm actually collects against what it originally worked — is relatively unchanged since 2023 for midsize firms. The Second Hundred had been sliding, but since the midpoint of 2025 they have recovered a full percentage point. Only the Am Law 100, the segment pressing hardest, has given up any real ground, and 1.5 percentage points against 9.7% rate growth in 2026 alone is not meaningful resistance.

With so little client pushback, it would be easy to conclude that all must be rosy for firms at the current time — and to some extent, you’d be correct.

Firm profitability is extremely strong halfway through the year, but it’s a strength blunted somewhat by the vast spending firms are doing, mostly on technology and talent, even as higher levels of rates and demand are rolling in the door.

‍

Overall expense growth for firms sits at a historically elevated level, overshadowed only by the pandemic-era super spending of 2022. Direct expenses⁵ alone are up 5.3% per full-time equivalent (FTE) employees over June 2025 levels, while overhead expenses are marginally stronger at 5.5%. Both categories are running hot compared to inflation, as they have been over the last couple of years while firms have embroiled themselves in competition over talent and infrastructure. These growth rates pale in comparison, however, to that of technology and knowledge management (KM), which clocked 10.3% growth per lawyer through June 2026, nearly double the pace of direct and overhead expenses overall.

Indeed, law firms are not in the business of spending money out of a sense of amusement, and this stretch is no different. The entire industry is locked in a technological adoption phase that is more expensive and more fundamental than any in its history.

What is the market saying?

Which brings us to the question everyone actually wants answered. Firms are spending much more on technology per lawyer than they were even a year ago, on top of direct expense growth that’s already running hot. So, what is that doing to the market?

The honest answer is that it depends entirely on how far back you stand.

Up close, the evidence is everywhere, and much of it is impressive. Ask around and you will hear about the partner who got a dozen hours of their week back, the practice group whose realization moved several points in a single year, the workflow that used to take three associates and now takes one. These accounts are specific, usually credible, and there are a great many of them.

Yet, a collection of anecdotal examples is not a market. Zoom out from the practice group to the firm and the effect stops being as visible. And this has become a common enough complaint among the people responsible for measuring such things that it deserves to be taken at face value. At the individual level you can find the AI-driven time savings, but at the firm level the signal gets lost in the noise.

The reason is not hard to find. Our 2026 Future of Professionals report found that 69% of law firm professionals say they now use AI tools at least several times a week, which is remarkable penetration for a three-year-old technology. However, 92% of those same professionals say their firms are falling short of what the technology could deliver; and our AI in Professional Services rereport found that 85% say their firm either isn’t collecting AI return on investment (ROI) metrics at all or that they don’t know whether it is. Meanwhile, only 41% of professionals within law firms say their legal teams are using GenAI, up from 28% the year prior, which represents strong growth although is still a minority of the market.

Individual lawyers have adopted AI at a pace the industry has not seen in a generation. Law firms, however, have not yet rebuilt themselves around it.

Read those numbers together and the picture is clear enough. Individual lawyers have adopted AI at a pace the industry has not seen in a generation. Law firms, however, have not yet rebuilt themselves around it. The tools are in use, but the workflows, the pricing protocols, the staffing models, and the measurement apparatus mostly are not, and a firm cannot measure what it has not yet systematized.

We made several attempts to find an answer to this incongruity, and each one came back empty or inconclusive — not because the data was thin, but because the effect, if it exists, is not where conventional tests would look for it. The simplest test was a tripwire: if AI were doing the work associates used to do, associate productivity should show it. Instead, associates worked more hours in total while simultaneously gaining in productivity.

We then stepped back and sorted firms by how much they spend on technology and KM per lawyer, comparing the one-third of firms spending the most against the one-third spending the least. The raw gap was striking: nearly three percentage points more worked rate growth for the top spenders, and double the growth in fees worked per lawyer. However, almost all of it vanished once we controlled for the inconvenient fact that we had largely been ranking firms by how much money they had. Adjusted for segment composition, both groups landed within the range of ordinary variation for peer group comparisons.

Finally, we stopped sorting firms and sorted the work instead, splitting the practice areas by how well current models perform on the tasks that fill them rather than by anyone’s estimate of what ought to be automatable. If AI were moving this market, the effect should have arrived sorted. It didn’t. The practices most exposed to what AI can actually do don’t look compressed; they just look busy. All three tests shared the same flaw: each arrived with a theory about where the effect should be. So, we inverted the procedure and asked the data to tell us when law firm economics changed shape.

The breaking point

First, we take inflation-adjusted technology and KM spend per lawyer, month by month, and draw the straight line that best fits its growth rate. Then ask one question: Is there a point where a single line stops working? One in which the trend afterward is different enough from the trend before that you need two lines to describe it rather than one? That point is a breakpoint.

The test carries no assumptions about either law firms or AI. It only knows whether the level or slope of a series' trajectory changed at a statistically identifiable point in time.

We ran two versions of the test, built to follow different logic and, presumably, to reach different results. The first was told to find exactly two breaks; the second could find as many as the data would support, but every break carried a small penalty, so the model had to earn each one rather than chase monthly noise. Both returned the same answer: law firm spending behavior fundamentally changed twice since 2022: First in November 2023, and again in April 2025.

‍

Both of these dates are historically significant. November 2023 is the month Microsoft Copilot reached general availability and firms could begin purchasing business-oriented GenAI products. In other words, November 2023 is the month this technology stopped being a demonstration and became a line item.

April 2025 was Liberation Day, launching the tariff regime that followed — and from that month forward, firms accelerated their technology spending further. That second date does two things at once. It is the closest thing to a sanity check this method can offer: asked when the legal market changed, an algorithm with no opinions about law firms or AI surfaced an event nobody disputes. It is also the muddying effect which helped obscure the findings in our previous tests, so seeing it surface here is expected.

Then, for each law firm segment, we extrapolate its own pre-November-2023 trend forward, treat that as the counterfactual, and measure the gap against what firms actually achieved across technology and KM overhead expense growth per lawyer, revenue, and profit. And since we kept every figure as per lawyer, that meant none of it is a story about firms simply getting bigger.

‍

The data shows that the entire industry shifted. Technology and KM overhead stepped up 0.70 points at the break and kept widening at 0.24 points a month — representing not a purchase, but the beginning of an arms race that’s only gotten more pitched. Revenue growth jumped 4.10 points above its expected path and has effectively stayed there, narrowing by 0.06 of a point monthly. Profit jumped hardest: 9.44 percentage points per lawyer.

Then consider what that number was being asked to beat. Indeed, 2022 was a poor year for profitability — profit per lawyer fell against 2021 — and the rebound that followed was steep. A trend fitted to a contraction and a sharp recovery has an aggressive slope, and extrapolating it forward demands not merely strong profit growth but continuously accelerating profit growth, indefinitely.

Yet, law firms beat that for the better part of two years. They now sit marginally below it, decaying at 0.46 points a month, while still posting profit growth per lawyer above 5%. The honest reading is not that firms fell short. It is that they nearly kept pace with a benchmark that was close to unbeatable by construction.

‍

Further, the segments showed significant divergence both in how spending translated into profit and in how long the initial post-breakpoint gain in profit and revenue held after the initial jump. Midsize and Second Hundred firms stepped up and stayed up, the latter posting the largest projected gain of the three segments while its own spending trend barely deviated from expectation.

The Am Law 100 — the segment spending most aggressively per lawyer — stepped up the least and is the only one whose profit deceleration is statistically significant. However, that is less damning than it appears. Am Law 100 firms had the strongest 2022 of any segment and rebounded the hardest from it, which hands them the steepest pre-break trend in the market. They also remain, throughout, the most profitable segment per lawyer in the industry and the fastest growing. Falling behind your own previous trend is not the same as falling behind in totality.

What this does and does not prove

Before we get to the final conclusions, we should be clear what the above analysis does not show. First, it does not isolate AI: the break is in all technology and KM spending, of which GenAI is one component and never the whole.

Further, the second break contaminates the first. Our counterfactual runs from November 2023 forward, which means the window meant to reveal a technology effect contains a trade war, the same geopolitical development that’s muddied the waters on determining clear market direction. This design cannot cleanly separate the two. What we can say is directional: April 2025 cannot explain a step-up that happened seventeen months before it, but it can very plausibly explain why that step-up widened and held. So, if we’re trying to determine how the legal market structurally changed via a method with no theory or understanding of events beyond the metrics given, it has returned two dates. One is the month legal AI went on sale; and the other is the month the trade war significantly escalated.

Law firms have seen a meaningful surge in profitability that began before the tumultuous trade conflict, and that continues to this day.

This is the ultimate takeaway. Law firms’ current prosperity — seen with record high prices that are increasing as fast as ever, with profitability soaring despite stiff competition over talent and technology — really has two parents. First, the trade war sent clients scurrying to their law firms to cover a variety of legal matter needs. Indeed, that boost in client demand, and clients’ willingness or need to bear ever-higher prices, swamped law firms’ higher expenses.

Second, AI itself is driving demand as clients reach out to their law firms for a variety of AI-related legal matters — from licensing to data center financing, to privacy and liability issues. The statistics still cannot show that firms’ own investments in legal technology, and GenAI in particular, have generated provable ROI at the macroeconomic level. What they do show is that the surge in profitability began seventeen months before the trade war escalated, which is why we read the November 2023 break as the technology-driven one. This means that the major impact of AI on law firms may be demand — even though the impact on firm spending and efficiency is likely to continue to gain traction.

In short, law firms have seen a meaningful surge in profitability that began before the tumultuous trade conflict, and that continues to this day.

Three years in

It's worth remembering how little time has actually passed since law firms got their hands on this technology for the first time in any serious way — just three years ago. November 2023, and nothing since email has moved through this industry this fast. And email had the advantage of not requiring anyone to rethink how work gets priced, staffed, or supervised. Three years is not long enough to rebuild a firm around a technology, however, and as the adoption figures earlier in this report show, most firms have not.

What this year’s Law Firm Rates Report does show, and shows clearly, is that AI is already reshaping law firm economics — just not, at least for now, from the inside.

The analysis we conducted found a genuine structural shift in firm performance beginning in November 2023, a step-up of 4.1 percentage points in revenue growth and 9.4 points in profit per lawyer above their expected paths, and it has largely held.

Firms did not earn that by automating their own document review. They earned it because the AI economy generates legal work through a variety of matters, from capital raisings and financings, to licensing disputes and the IP fights. Our research shows that, at least immediately, the impact of the AI boom is reaching law firms as demand, not as efficiency — even though that may not be where the impact was most expected. Indeed, between AI-related legal demand and a continuing trade war that sent clients speeding to their legal advisors, demand is the channel our data can actually account for in 2026. Whether that continues, or the AI impact shifts to the other side of the ledger, pushing efficiencies and rate pressure from clients, is the key question for 2027.

Law firms are not living with the AI impact because of what it has done to their work; rather, they are living with it because of what AI is doing to their clients' world and to the volume of exposure that it has generated.

Admittedly, we expected that latter result at the onset of our research. We thought we’d find advanced technology at work inside the firm, with hours falling, output rising, and the efficiency gain showing up in the rate — upward, we thought, since a firm that gets more done inside an hour has not made that hour cheaper, it has made it worth more.

Instead, we needed to find the answer by looking into when law firm economics changed shape and were told that the month legal GenAI went on sale was the linchpin.

Simply, firms are not living with the AI impact because of what it has done to their work; rather, they are living with it because of what AI is doing to their clients' world and to the volume of exposure that it has generated. And that means, of course, this era may have a second half, and firms should not expect it to resemble the first. The first half pays firms to handle the disruption by raising the capital, papering the deals, fighting over the rights, and answering the regulators. That work is a byproduct of a technology being deployed at speed by everyone else, and it is generous, but it is finite in a way nobody can date.

The second half is the part in which the technology reaches the work itself, and where the questions this report could not answer become the only questions that matter. What does a leveraged pyramid look like, and how is its work priced, when much of what the bottom rungs have always done is automated away? What happens to a pricing model built on a billable hour when the hour itself stops being the unit of production?

Although nothing in our data says that the second half has begun for law firms, nothing says it won't, either. So, before that second part of the AI Age arrives, law firms would be wise to use this first half— the one which is paying handsomely for now — to build in their AI resiliency for the coming second half.

For example, here are several actions firms should take now and what they should be on the lookout for as they enter this second half.

‍

Indeed, while no one can predict exactly when the second half will begin, acting and being aware now is the best use anyone will ever find for a golden age, and it is the one thing that cannot be bought back once the age is over.