Can Legal AI (Ever) IPO?
Well, it depends. Lets layout the conditions needed pave a realistic path to actually printing money. Previously, I stated most top legal tech companies can not IPO. Market conditions have not improved since. Nevertheless, any crisis offers entrepreneurs opportunities to capitalize.
From Chaos 2 Capital
The reasons a legal tech venture goes from startup to successful exit are laid out in Anatomy of a Legal Acquisition. TL;DR: the most upside is gained by not raising a lot. In a post on Artificial Lawyer, I shared data comparing valuations vs exit value numbers. Why not raise a lot or even better, bootstrap? Because that makes you and your company affordable. Once a company has 'borrowed' over certain amount, options for a lucrative payday begin to evaporate. The only recourse is to offer your company to the public through intial public offering (IPO). I would not bet on support from Private Equity. Just ask Relativity. Also, remember when Docusign tried to go private? Going through a Special Acquisition Company (SPAC) is no guarantee either. SPACed companies like FiscalNote have not given me much hope based on their stock price history.
Surely there have been successful legal tech IPOs? Nope, not really. In 2023 we analyzed 17 public legal tech companies and all except the big 3 did poorly. Currently, even the three centuries‑old companies dominating legal technology have lost more than half their value, with no signs of recovering it. Worse, their decline started in a very peculiar point, way before the infamous Anthropic Legal plugin dropped. Now these unprecedented low public valuations sit in stark contrast to late‑stage private legal tech valuations.
Fine, but this would not stop current legal AI companies from seeking an IPO, right? Relativity, a profitable legal data company, filed confidentially for its IPO on March 19. It would be the first legal tech company to go public since 2021. SpaceX filed 13 days later, on April 1, and is already trading. Relativity is not. Relativity still has no public prospectus, no price, no ticker.
Recap: If you are not cheap enough to be acquired, your plans to cash out reduce to zero. Unless you look like Palantir.
AI Supply vs AI Demand
As demand for AI compute grows, it has also drained the capital supply needed to fund that growth. Except for Apple, the rest of the ‘Magnificent Seven’ are borrowing because they have no more free cash left to spend. The growth in demand for AI compute correlates with widespread use of AI. While AI use is popular with most, only a few sectors can afford to pay. The sector that has a lot of cash to spend seems to be law firms. Boy, spending on legal AI is crazy if you believe the ARR announcements from major players. Unfortunately, not everyone is so lucky and there are reports of layoffs by legal AI vendors. In other words, both legal AI budgets and AI compute spending are hitting real limits.
For argument's sake, let's say supply of AI stays abundant, as it currently still is. Where is the scarcity? Well, accuracy and security are still in short supply in the use of AI. That is why Palantir set record-breaking 85% revenue surge. Better yet, they are arguably the only profitable software company selling AI. Since they are a public company, we can verify their reported numbers. Wonder why? Palantir sells on-premise AI (security) and an ontology (accuracy). Hence, I unretired from writing these analysis to explain Kirkland & Ellis's $500 million bet using Palantir tech.
Another thing in short supply is affordable AI. I just rebuilt Legalcomplex on roughly 5 billion tokens. But the kicker is that I only paid €120 for services that would cost around $10,300. Which is why no one is making a profit except Palantir. Because everyone's AI use is subsidized. Especially frontier AI use. So there is demand for accurate, affordable AI and Autonomy. No one is able to supply, and if they claim they can, they are lying. Lying because the constraints are physical not digital: Chips, Electricity, Water and Real Estate.
Point: accurate, affordable AI that offers you autonomy is in short supply, everything else will likely go to zero due to abundance.
Humand-verified vs AI-Automation
Obviously, only hardware companies offering physical solutions for the AI are capitalizing. Software companies that support can accurate, affordable AI and Autonomy have slim shot at profiting like Palantir. Examples are suppliers of unique data like Reddit user generate answers. as a matter of fact yet, the only billionaires being created by the AI boom supply human-verified AI data. While human-verified data is proprietary, not all proprietary data is valuable. The data that is not valuable is government-produced legal data like legislation and case law. Although legal data is protected by copyright, governments are legally obligated to provide citizens with access to the law. This is contrary to privately produced legal data, see Kirkland & Ellis.
The problem with privately produced legal data is that it is niche and thus only valuable under specific conditions. Same goes for automating legal processes. The key to identifying valuable opportunities is distinguishing volume from value. The acquisition of Twitter constitutes a high-value, low-volume legal process, as do fundraising and major corporate litigation. Given these high stakes, those processes remain human-bound despite being highly automatable. Not because humans are more accurate, far from it. It's because they are accountable. Last year, I gave a presentation in Madrid arguing that firms will be the ultimate winners in legal AI for one simple reason: accountability.
Although nearly all legal processes lend themselves to AI processing, only a subset requires human verification. Identifying this distinction demands the skill to determine which processes hold value. Legal practitioners may claim human involvement is essential in all legal processes, yet judicial recognition of e-discovery indicates otherwise. Despite this major legal victory, it has not translated into business wins for legal tech vendors. As I said: no successful public legal tech companies yet. Maybe Relativity? Smirk.
IPO: To all legal tech companies: if you’ve raised $100 million or less, think very hard before raising again. Might still have acquisition, PE or SPAC exit option available. If you are above $100 Million than showing your shares onto the public may be your only option. In that case, you need growth, crazy growth. You can only get that if you have no competitors in your market.
Based on all of the above, the only path to IPO is building something that is genuinely scarce: accuracy, security, or autonomy.
Private Prologue
I used to write every month, but stopped for one reason: AI. I will start again to organize my thoughts and rationalize my reasons.
Sources linked inline. Legal-tech funding and M&A counts from Spark by Legalcomplex (legal segment). Figures dated to mid-2026. ~1,000 words, 5-minute read.
