Which legal AI startups are generating the most revenue today?
In our updated market reports, you will find everything you need
SUMMARY
Harvey is clearly generating the most revenue among AI-native legal startups today, at roughly $350 million in ARR, with Legora and EvenUp forming the next tier.
The first thing to get right is the category. If established legal-software companies that later became heavily AI-driven are included, Clio moves to the top at more than $500 million ARR, while Filevine and Ironclad also enter the upper end of the ranking.
Legal AI has already produced three AI-native businesses at or around nine-figure recurring revenue: Harvey, Legora and EvenUp. That is a much more mature revenue picture than the funding headlines alone suggest.
Harvey’s lead is not mainly about having the most customer logos. It comes from selling very large deployments into global law firms and enterprises, which pushes its crude average revenue per customer above $230,000.
Legora is the only AI-native challenger close enough to make the race interesting. Its disclosed floor is above $100 million ARR, and the best current estimates place it near $150 million, still well behind Harvey but growing unusually fast.
EvenUp shows that Big Law is not the only legal niche capable of supporting a $100 million AI company. Personal-injury firms can also support very large software and workflow contracts, including at least one disclosed account spending more than $4 million a year.
LegalOn and Luminance sit in a different but important tier. LegalOn has a hard disclosed ARR floor around $67 million and more than 8,500 customers, while Luminance looks like a $60 million-plus business with at least one eight-figure enterprise contract.
Below the leaders, fake precision becomes a real problem. Spellbook is targeting $100 million ARR, Solve Intelligence only says eight-figure ARR, and Supio has disclosed 17-fold growth without the starting base, so ranking them to the nearest million would be nonsense.
The market is already concentrating, but not around one legal workflow. Big Law, personal injury, corporate legal departments, patents, contracts and litigation are all producing meaningful revenue, which leaves room for specialists even as the biggest platforms pull away.
The main competitive threat is not that foundation-model companies simply erase legal AI startups. The stronger pattern so far is a layered market: foundation models underneath, specialist legal AI in the middle, and firm-specific or enterprise workflows on top.
The revenue leaderboard and the profitability leaderboard are also different. Harvey leads AI-native revenue while remaining unprofitable, whereas Clio, Solve Intelligence and Darrow show that smaller or more mature legal-tech businesses can already operate profitably.
Which legal AI startups are generating the most revenue today?
What actually counts as a legal AI startup?
For this ranking, we count private companies whose core business depends on using AI for legal work, which puts Harvey, Legora, EvenUp, LegalOn, Luminance, Spellbook and several newer specialists in the main group.
The definition matters more than it might seem. Clio now calls itself a legal AI platform and has passed $500 million in ARR. Filevine is estimated at roughly $250 million ARR, while Ironclad has officially passed $200 million. All three built large legal-software businesses years before the current generative-AI wave.
Including them would answer a slightly different question: which private legal-software companies using AI make the most money? That ranking starts with Clio.
We are trying to identify the companies whose commercial rise is actually tied to AI. On that basis, Harvey leads comfortably.
| Company | Latest useful revenue figure | How we treat it |
|---|---|---|
| Clio | $500M+ ARR | Established legal software turned AI platform |
| Harvey | ~$350M ARR | AI-native legal startup |
| Filevine | ~$250M ARR estimate | Established legal platform with major AI push |
| Ironclad | $200M+ ARR | Established contract platform with major AI push |
| Legora | $100M+ disclosed; ~$150M estimated | AI-native legal startup |
If you want more recent data on this point, please see our latest legal AI market report.
Why is it so hard to rank legal AI startups by revenue?
Legal AI revenue is surprisingly difficult to rank today because companies disclose different numbers, at different times, under different definitions.
Harvey provides a good example. Recent reporting puts the company at roughly $350 million in ARR. Previous disclosures had Harvey around $195 million at the end of last year and above $100 million only a few months before that. The direction is clear even though individual reports have sometimes alternated between ARR and annualized revenue.
Legora officially announced more than $100 million ARR after reaching the milestone in just 18 months. More recent private-market estimates put Legora near $150 million. We can confidently say that Legora has a nine-figure recurring-revenue business; the exact current number carries less certainty than the official $100 million floor.
The problem gets worse farther down the ranking. Solve Intelligence describes itself as having eight-figure ARR, which could mean anywhere from $10 million to $99 million. Supio says ARR has increased 17-fold without disclosing the starting number. Luminance says global revenue doubled for a second consecutive year but leaves the absolute amount private.
So we can rank the top few companies quite confidently. Below them, revenue bands are more honest than fake precision.
Is Harvey the highest-revenue AI-native legal startup today?
Yes. Harvey is currently the clear revenue leader among AI-native legal startups, at roughly $350 million ARR.
That number has just received another useful check. In a recent interview with The Times, Harvey co-founder Winston Weinberg put the company at $350 million in annual recurring revenue, more than 200,000 lawyers using the product and roughly 1,200 employees. Sacra independently estimates Harvey at the same $350 million level.
The growth curve is extreme. Harvey was around $50 million ARR at the end of 2024, crossed $100 million during 2025 and finished that year around $195 million. Reaching roughly $350 million since then means the business has grown about sevenfold from its late-2024 level and by roughly 80% from the end of last year.
Usage has expanded alongside revenue. Current estimates put Harvey at around 1,500 customer organizations across about 60 countries. Sacra says weekly active users have increased fourfold year over year, monthly queries 5.5 times and active files to nearly 10 million.
That is much more convincing than another giant financing round. Harvey has built a several-hundred-million-dollar recurring software business from legal AI in roughly four years.
Is Legora actually catching Harvey?
Legora is currently the only AI-native legal startup close enough to Harvey to make the comparison interesting, but Harvey still generates well over twice as much recurring revenue.
Legora officially crossed $100 million ARR after just 18 months of general availability. Bessemer described it as the fastest enterprise software company it had seen reach that milestone. More recent estimates place Legora around $150 million ARR.
The acceleration is remarkable. Sacra estimates Legora at roughly $3 million ARR at the end of 2024, around $50 million one year later and roughly $150 million by the middle of this year. Even if those estimates are only directionally correct, Legora has moved from a small startup to one of legal software's biggest independent businesses extraordinarily quickly.
Customer growth supports the revenue story. Legora had around 250 customers in the spring of last year, passed 400 later that year and now serves more than 1,000 organizations across over 50 markets. Clients include White & Case, Linklaters, Cleary Gottlieb, Barclays and Blackstone.
Harvey still has a roughly $200 million revenue advantage using the best current estimates. Legora's growth rate makes it the obvious challenger, but closing a gap this large while Harvey itself keeps growing will be difficult.
If you want more recent data on this point, please see our latest legal AI market report.
Has EvenUp already become a $100 million legal AI company?
Yes. The strongest recent evidence now puts EvenUp at roughly $100 million ARR, making it one of the three biggest AI-native legal startups we can identify.
EvenUp has taken a completely different route from Harvey and Legora. The company concentrates on personal-injury law, where AI can organize medical records, prepare demand packages, evaluate cases and help law firms process far more claims.
That narrow focus has produced surprisingly large contracts. EvenUp's founders told Fortune that their largest customer spends more than $4 million a year. The company already serves more than 2,000 personal-injury firms, giving it both substantial customer volume and unusually high spending among its largest accounts.
The historical numbers make the current $100 million figure believable. Reporting in 2024 put EvenUp at roughly $35 million in annualized revenue with expectations of passing $50 million. The company later said ARR had doubled year over year. Recent revenue tracking now puts it at the $100 million mark.
That places EvenUp behind Harvey and probably Legora, while clearly separating it from most of the remaining AI-native legal startups.
Is LegalOn bigger than most people realize?
Yes. LegalOn is already one of the world's biggest AI-focused legal software companies, with a hard disclosed ARR floor of roughly $67 million and more than 8,500 corporate customers today.
LegalOn passed ¥10 billion in ARR, approximately $67 million at the time of the announcement, after six and a half years in market. The company began with AI contract review and has expanded into contract management, matter management, compliance and broader legal workflows.
The customer base is the part that stands out. LegalOn said earlier this year that more than 8,500 companies use its products globally, up from more than 7,000 when it disclosed the ¥10 billion ARR milestone. More than 30% of listed companies in Japan use LegalOn.
This gives LegalOn a commercial profile very different from Harvey's. Harvey earns far more revenue from far fewer organizations, while LegalOn has built much broader penetration, especially across Japanese corporate legal departments.
We have no newer absolute ARR disclosure that justifies projecting LegalOn far beyond $67 million. What we can say confidently is that LegalOn has already reached a scale that most legal AI startups have yet to approach.
Is Spellbook already generating $100 million a year?
No. Spellbook is growing toward $100 million ARR, but the company has presented that number as a target rather than a revenue milestone already reached.
Spellbook tripled revenue during 2025 and grew from roughly 4,000 customers last year to more than 4,500 law firms and in-house teams across around 80 countries. CEO Scott Stevenson has said the company's current trajectory could take it to $100 million ARR by the end of the year.
That makes Spellbook one of the more interesting companies just below the obvious leaders. Its customer base already exceeds Harvey's by a wide margin because Spellbook sells much more heavily into smaller legal teams and smaller firms.
The economics are therefore different. Harvey can generate several hundred million dollars from roughly 1,500 organizations because some deployments cover hundreds or thousands of lawyers. Spellbook spreads revenue across thousands of smaller accounts.
Until Spellbook actually reports the $100 million milestone, we should keep it below companies that have already crossed that threshold.
How big is Luminance today?
Luminance looks like a roughly $60 million-plus legal AI business today, although its exact ARR deserves more caution than Harvey's, Legora's or LegalOn's figures.
Sacra estimated Luminance at approximately $60 million ARR at the end of 2025, double the roughly $30 million estimate from one year earlier. Luminance itself has said global revenue doubled for a second consecutive year and that North American revenue increased 127%.
One number jumps out: the company signed its first eight-figure enterprise contract. That means at least one customer relationship is worth $10 million or more, an unusually large account for legal software.
Luminance also says its AI has now analyzed more than 220 million documents. Its business has expanded from due diligence and contract analysis into a broader enterprise legal platform used for negotiation, workflow and institutional knowledge.
The $60 million estimate is already several months old, so Luminance could be meaningfully above it now. Without a newer absolute disclosure, we would keep it in the $60 million-plus tier rather than manufacture a current figure.
Which smaller legal AI startups are already making serious revenue?
Eudia, Darrow, GC AI, Solve Intelligence and Lawhive have all reached meaningful eight-figure scale, while Supio is growing fast enough to watch closely even though its absolute revenue remains undisclosed.
Eudia gives us one of the cleanest numbers. Founder Omar Haroun said the company went from roughly $2 million to $20 million ARR in twelve months. Eudia targets large corporate legal departments and combines AI with legal services, which lets it compete for spending that would traditionally go to outside counsel.
GC AI has followed a similar rise in the in-house market. The company publicly crossed $10 million ARR late last year, and Sacra now estimates it at roughly $20 million. Its estimated customer base of around 1,700 legal teams shows how quickly relatively young legal AI products can reach broad corporate adoption.
Solve Intelligence has disclosed eight-figure ARR and is growing 20% to 30% month over month according to recent hiring material. A spring interview with co-founder Chris Parsonson put the company around $13 million ARR after starting from zero less than two years earlier. Solve focuses specifically on patents and now serves hundreds of IP teams.
Darrow is harder to rank because its last reliable revenue figure is older. The company generated $26 million in 2024 and forecast more than $50 million the following year, with management targeting as much as $100 million in revenue this year. We found no fresh evidence that the $100 million target has actually been reached, so we leave it out of the nine-figure group.
Lawhive reported more than $35 million in annual revenue earlier this year after growing sevenfold. Its model includes legal services delivered through lawyers using AI, so every dollar of Lawhive revenue is economically different from pure software ARR.
Supio recently reported 17-fold ARR growth since emerging from stealth and said its customer base doubled during the first half of this year. The missing absolute revenue number prevents a serious ranking. Seventeen times $1 million and seventeen times $5 million tell completely different stories.
Why can Harvey make more money with fewer customers?
Harvey makes far more revenue per customer because it sells large enterprise deployments into global law firms, financial institutions and corporate legal departments.
Using current figures, Harvey has roughly 1,500 customer organizations and about $350 million ARR. That works out to more than $230,000 of ARR per organization as a crude average.
LegalOn has more than 8,500 companies and a last disclosed ARR figure around $67 million. GC AI has roughly 1,700 customers and an estimated $20 million ARR. Spellbook has more than 4,500 customers while still working toward the $100 million mark.
These averages are imperfect because customer definitions and reporting periods differ, but the difference is too large to ignore. Harvey has built a business around organizations that can deploy its product across hundreds or thousands of lawyers.
The latest user numbers reinforce the point. Harvey now says more than 200,000 lawyers use its platform. Dividing revenue by customer count therefore hides the real engine: each enterprise logo can represent an enormous number of paid users and workflows.
This explains much of Harvey's revenue lead today. Its advantage comes from combining rapid customer acquisition with unusually large enterprise accounts.
Is Big Law still where legal AI makes the most money?
Big Law has produced the biggest individual legal AI winner so far, but personal injury and corporate legal departments are already supporting businesses worth tens or hundreds of millions of dollars.
Harvey and Legora show why large law firms were such attractive early customers. A single global firm can roll software out to thousands of lawyers, while legal work carries enough economic value to support high software spending.
EvenUp proves that personal injury can produce similar economics through a different route. Its largest disclosed account spends more than $4 million annually, and the company has reached roughly $100 million ARR while specializing almost entirely in plaintiff work.
Corporate legal departments form another large market. LegalOn has more than 8,500 companies. Eudia has grown to roughly $20 million ARR by targeting enterprise legal spending. GC AI is around the same level with a much broader base of in-house teams. Luminance has secured at least one eight-figure enterprise deal.
So the revenue opportunity has already spread well beyond Big Law. The legal AI companies making serious money tend to own expensive workflows where saving lawyer time or reducing external legal spend has an obvious dollar value.
If you want more recent data on this point, please see our latest legal AI market report.
Are older legal-tech companies still making more money than Harvey?
Clio still generates more recurring revenue than Harvey, while Filevine and Ironclad are already in the same broad revenue league as the AI-native leaders.
Clio passed $500 million ARR earlier this year and says it remains profitable. The interesting part is the trajectory: Clio was above $200 million ARR around the middle of 2024, reached roughly $400 million late last year and then moved beyond $500 million as AI became more central to its platform and the company integrated legal-research provider vLex.
Filevine is currently estimated at roughly $250 million ARR, up from about $130 million in 2024. Its advantage is distribution. Filevine can push products such as LOIS, its AI legal assistant and research system, into an existing case-management customer base.
Ironclad officially crossed $200 million ARR earlier this year. It has since expanded further into AI assistants and contracting agents while partnering with Legora for in-house legal workflows.
The installed-base advantage is becoming hard to ignore. AI-native startups can grow much faster from zero, while established legal platforms already control daily workflows, documents and customer relationships.
Still, Harvey has already passed Filevine and Ironclad on the best current revenue evidence. Only Clio clearly remains larger.
Are the highest-revenue legal AI startups profitable yet?
Revenue leadership has not yet translated into profitability for Harvey, while several smaller or older legal AI businesses already say they make money.
The latest profile of Harvey puts the company at roughly $350 million ARR, around 1,200 employees and still unprofitable. That tells us how aggressively Harvey is spending on product development, enterprise sales and international expansion.
Clio offers the clearest contrast. It says it has passed $500 million ARR while remaining profitable. Clio benefits from an 18-year-old software business, a huge installed customer base and a much more mature operating model.
Solve Intelligence says it is profitable at eight-figure ARR. Darrow has said it has been cash-flow positive since 2023. Their revenue is much smaller than Harvey's, but they show that legal AI can already support profitable companies.
EvenUp has historically carried more human review and operational work than a conventional SaaS vendor, which can reduce gross margins. Eudia and Lawhive also blend software with legal services to varying degrees.
Harvey wins the AI-native revenue race. Profitability is a separate leaderboard, and Harvey does not lead that one today.
Is legal AI revenue already concentrating around a few winners?
Yes. Revenue is already becoming concentrated around a small group, with Harvey far ahead and a noticeable drop after the first three or four companies.
Harvey contributes roughly $350 million ARR by itself. Legora is somewhere above $100 million and plausibly around $150 million. EvenUp is now around $100 million. LegalOn's last hard disclosure was approximately $67 million, while Luminance is estimated around $60 million.
Using only those conservative figures gives us roughly $700 million of recurring or annualized revenue across five AI-focused legal companies. Spellbook, Eudia, GC AI, Solve Intelligence, Darrow, Supio and others push the visible market higher.
The distribution is steep. Harvey alone currently generates roughly as much recurring revenue as Legora, EvenUp and one smaller eight-figure company combined.
There is still plenty of room for specialists because legal work is fragmented across contracts, patents, litigation, personal injury, research and corporate workflows. But the idea that hundreds of generic legal copilots will divide the market evenly already looks implausible.
If you want more recent data on this point, please see our latest legal AI market report.
Can OpenAI, Anthropic and the legal giants crush Harvey and Legora?
They can put real pressure on Harvey and Legora, but crushing them is not what the revenue data shows so far.
The competitive environment has become much tougher lately. General AI companies are building better research, document analysis and agent capabilities. Thomson Reuters is expanding CoCounsel. LexisNexis has decades of proprietary legal data and embedded lawyer relationships. Google Cloud recently launched Gemini Enterprise for Legal with specialist partners including Solve Intelligence.
Large law firms are also building more technology internally. Recent Financial Times reporting found that firms such as Kirkland & Ellis and Freshfields are investing heavily in custom AI, with Kirkland committing $500 million to its own technology program.
Yet Harvey has gone from roughly $195 million to $350 million ARR while these alternatives have become stronger. Legora crossed $100 million during the same period. Solve Intelligence has joined Google's legal AI ecosystem rather than disappearing beneath it. Ironclad and Legora have chosen to integrate their products with each other.
The market increasingly looks like several layers working together: foundation models underneath, specialized legal AI in the middle and law-firm or enterprise workflows on top. The startups that own valuable workflow and customer relationships still have plenty of room.
Which legal AI startups are generating the most revenue today?
Harvey is clearly the highest-revenue AI-native legal startup today, followed by Legora and EvenUp; LegalOn and Luminance form the next credible revenue tier.
Harvey's position is the easiest call in the entire analysis. Two fresh sources now put it around $350 million ARR, after roughly $195 million at the end of last year. We found no AI-native legal competitor with a current revenue figure close enough to overturn that lead.
Legora comes second on the best available evidence. Its company-disclosed floor is above $100 million ARR, and current estimates put it around $150 million. EvenUp has now reached roughly $100 million ARR, giving legal AI three AI-native companies at or around nine-figure recurring revenue.
LegalOn follows with a hard disclosed floor of about $67 million and continued customer growth since that milestone. Luminance was estimated around $60 million at the end of last year and has continued expanding. Spellbook appears to sit somewhere below $100 million today while targeting that threshold.
After that, confidence drops. Lawhive has more than $35 million in annual revenue, though its model includes legal services. Darrow's last hard number was $26 million. Eudia and GC AI are around $20 million. Solve Intelligence is in eight-figure ARR and growing quickly. Supio has disclosed extraordinary percentage growth without revealing enough to rank it properly.
If we broaden the question to include older legal-tech startups that have become heavily AI-driven, Clio takes first place at more than $500 million ARR. Filevine at roughly $250 million and Ironclad above $200 million also become major contenders.
For the AI-native category, the ranking today is clear enough. Harvey has separated from the field, Legora is the strongest challenger, and EvenUp has quietly built another $100 million business in a very different corner of law. The market has moved much further than the funding headlines suggest: several legal AI startups now generate tens of millions of dollars a year, three have reached roughly $100 million or more, and Harvey is already operating at a scale once associated with mature enterprise-software companies.
| Rank | AI-native legal company | Best current revenue evidence | Our confidence |
|---|---|---|---|
| 1 | Harvey | ~$350M ARR | Very high |
| 2 | Legora | $100M+ disclosed; ~$150M estimated | High |
| 3 | EvenUp | ~$100M ARR | Medium-high |
| 4 | LegalOn | ~$67M disclosed ARR floor | High |
| 5 | Luminance | ~$60M ARR estimate, with newer growth since | Medium |
| 6 | Spellbook | Below $100M; targeting $100M | Medium |
| 7 | Lawhive | $35M+ annual revenue | High, but services-heavy |
| 8–10 | Darrow / Eudia / GC AI | ~$20M–$30M range on latest hard evidence | Medium |
| Unranked | Solve Intelligence / Supio | Eight-figure ARR / 17x ARR growth | Insufficient precision |
If you want more recent data on this point, please see our latest legal AI market report.
OUR METHODOLOGY
This analysis asks which private legal AI companies are generating the most revenue today. There is no clean public dataset for that, so we built the ranking from the freshest available revenue disclosures, high-quality reporting, customer and user scale, growth milestones, enterprise contract size and operating evidence that could confirm or challenge the revenue picture.
We gave the most weight to achieved revenue, recent company disclosures and independently reported figures with clear sourcing. Targets stayed as targets, percentage growth without a disclosed base was not converted into artificial precision, and older figures were used mainly to understand trajectory rather than to pretend they were current.
Where several pieces of evidence pointed in the same direction, we used them together. That is especially important below the top few companies, where one company may disclose ARR, another annual revenue, another only a growth rate, and another nothing more precise than “eight-figure ARR.”
We also separated AI-native legal companies from established legal-software businesses that became heavily AI-driven later. That keeps the main ranking focused on companies whose commercial rise is tied directly to AI, while still showing how the answer changes when Clio, Filevine and Ironclad are included.
Key sources used for the ranking and cross-checks include The Times on Harvey’s current ARR and user scale, Harvey’s own revenue-growth disclosure, Legora’s $100 million-plus ARR announcement, Fortune on EvenUp’s customer economics, LegalOn’s ¥10 billion ARR disclosure, Luminance on revenue growth and its first eight-figure enterprise contract, Forbes on Spellbook’s growth and $100 million ARR target, Fortune on Lawhive’s annual revenue, Clio’s $500 million-plus ARR announcement, and Ironclad’s $200 million-plus ARR announcement.
For market structure and competitive context, we also used GC AI’s customer-scale disclosure, Google Cloud’s Gemini Enterprise for Legal announcement, the Legora-Ironclad integration announcement, and the Financial Times on Kirkland & Ellis’s $500 million AI technology commitment.
Related blog posts
