Who Wins the $1.1 Trillion Legal Market?

For decades, every services business has been told the same thing. You cannot scale like software. Your revenue is chained to your headcount. Your margins top out around 30% on a good day. Sell for a low multiple, if you can sell at all.

That just stopped being true, and this newsletter exists to track exactly how, where, and for whom.

Every two weeks I will write about the services economy as AI rebuilds it: the markets, the companies, and who captures the software economics as the work changes hands. I run Europe's first AI-native M&A advisory, so I am not writing about this shift from the outside. I am one of the entrants making the bet.

We start with the biggest arena of them all. Legal.

The shape of the battlefield

Two numbers frame everything. The global legal tech market is worth roughly $36 billion. The global legal services market is worth roughly $1.1 trillion. Every dollar spent on tools for lawyers is dwarfed thirty times over by dollars spent on the work itself.

For the past two years, most of the money and most of the headlines have crowded into the small number. Harvey has raised over $1 billion and is valued at $11 billion on roughly $190 million of ARR... a 58x multiple. Legora is at $5.6 billion after Nvidia and Atlassian joined its $600 million Series D, having tripled its valuation in five months and just crossed $100 million in ARR. Both sell software to law firms.

A 58x multiple on a company selling into a $36 billion market only makes sense if the real prize is the $1.1 trillion underneath it. Which is why the interesting question in legal is no longer "which tool wins". It is "who ends up owning the work". And right now there are three distinct routes onto that battlefield, each with a named player already committed.

Route one: the law firm becomes a tech company

In May, Kirkland & Ellis... the highest-grossing law firm in the world... committed $500 million over the next three to four years to build its own proprietary AI platform. Starting with $100 million this year. Designed on input from 250 of its own lawyers. Built by more than 180 technology professionals. And, crucially, never to be sold or licensed to anyone else.

Kirkland could have bought Harvey or Legora for a fraction of that. Their chairman's explanation of why they did not is the most important sentence in the story: off-the-shelf tools are "raising the floor for everyone". A tool anyone can buy makes everyone the same. If the firm across the street runs the same model on the same market knowledge, the AI does not widen your edge. It erases it.

What Kirkland is buying with $500 million is not intelligence. The models supply that. It is the encoding of their own tradecraft... the specific way their senior partners structure a fund... into a system that compounds instead of retiring.

Note what makes this route possible: Kirkland broke $10 billion in revenue last year, with profit per equity partner at $11.1 million. Half a billion dollars is under 5% of a single year's revenue. This route exists only at the very top of the market. Remember that, because it decides who loses.

Route two: the tech company becomes a law firm

On the opposite side, the entrants skipped the tool business entirely and became the firm.

Crosby is the sharpest example. It does not sell software to lawyers. It employs lawyers, arms them with its own AI, and sells the finished contract... priced per document, not per hour. In under a year out of stealth it has negotiated over $1 billion in contract value for around 100 clients including Cursor, Clay and Cognition, reviewed 13,000 contracts, and grown revenue roughly 400% since October. In March it raised a $60 million Series B. Lawhive runs the same play for consumers and SMEs: $116 million raised, $35 million ARR growing 7x, 500 lawyers across 35 US states. Norm AI raised $140 million, launched Norm Law for financial institutions, and hired the former chairman of Sidley Austin... the sixth-largest law firm in the world... to chair it.

And then there is the proof that the model survives contact with a courtroom. Garfield AI, the first SRA-regulated AI law firm in the UK, won a trial in the English courts this year. Its client paid £400 to recover £7,000. Look at how it won, because this is the operating manual for the whole category: the AI did the bounded, repeatable work... correspondence, filings, witness statements... and a human barrister did the advocacy. It is even barred from proposing case law, the single biggest hallucination risk, designed out from day one.

Compare that with the elite firms censured this year in London and New York for AI-hallucinated submissions. Same technology. Opposite results. The difference is not the model. It is that the AI-native firms rebuilt the work around what the model can and cannot do, while the incumbents pointed a clever tool at an unchanged process.

The honest caveat: every company on this route is winning a wedge, not the whole market. Contracts for fast-growing tech companies. Consumer claims. Small-value recoveries. None of them is doing bet-the-company litigation, and none will be soon. But wedges are how every disruption starts, and these wedges are compounding fast.

Route three: the tool that becomes the firm

Which brings us back to Harvey and Legora, and the most interesting strategic position on the board.

Here is their trap. They have the capability, the capital and the data to become the biggest law firms of the future. What they do not have is a business model that permits it, because becoming the firm means competing with the very customers paying for their licences. The moment they cross the line, the cancellations start.

But watch the M&A instead of the messaging, because the line is already being crossed... just not by them yet. Carta acquired Avantia and became a regulated law firm overnight. Lawhive opened a New York office with the stated purpose of acquiring US law firms and putting them on its platform. Meanwhile Legora's four acquisitions this year... Walter, Qura, Graceview, Cadastral... are all capability deals, the classic SaaS playbook of buying your way to a complete product. Two different M&A strategies that look identical on the deal page and are opposite bets on which category structure wins.

My read: the tools will cross eventually, because the maths forces them to. You cannot grow into an $11 billion valuation inside a $36 billion market. The only question is whether they cross before or after the AI-native firms and the Kirklands of the world have taken the ground they need.

The player not on the board

There is a fourth position in this market, and it belongs to the firms nobody is writing headlines about: the thousands of mid-market practices that cannot fund a Kirkland-scale build and are renting the same Harvey licence as the firm across the street.

They are about to discover the uncomfortable geometry of this fight. The ceiling is being defended by proprietary platforms they cannot afford. The floor is being taken by AI-native entrants with a cost structure they cannot match. And the tools they are renting raise the floor for everyone at once, which is precisely why they confer no edge. The middle of the legal market is not being attacked by anyone in particular. It is being squeezed by everyone at the same time.

My call

Rankings, so you can hold me to them.

The AI-native firms win the floor. High-volume, bounded, previously unservable work... the Crosby and Garfield territory... goes to them decisively. The economics are not close.

The Kirkland tier defends the ceiling. Bet-the-company work stays with elite human judgement, now compounding on proprietary platforms. The top ten firms get stronger, not weaker.

The tools face the hardest decade. Harvey and Legora are today's headline valuations, but they are squeezed between customers who fear them and a market too small for their price tags. Crossing into services is their only route to the valuation they already carry, and every quarter they wait, the natives take more of the ground.

The middle gets consolidated. The squeezed mid-market becomes acquisition inventory... for the natives, for the tools when they cross, and for the roll-up capital already circling professional services.

The pattern is not a legal story. It is the template. Every services vertical... accounting, insurance, banking, consulting... will resolve into the same four positions: incumbents who build, natives who enter, tools that must choose, and a middle that gets squeezed.

And the same fight is starting everywhere. In accounting, the incumbents are losing the people to fight it at all... 340,000 accountants gone in five years, and three-quarters of the remaining CPAs near retirement. In insurance, AI-native carriers and brokerages are already licensed and writing business. In consulting, the collision is only just beginning. That is what this newsletter is here to cover: wherever the fight is hottest.

If you run a services business and you are working out which of the four positions you occupy, that is exactly the conversation this newsletter is for. Subscribe, and I will see you in two weeks.