Harvey

1. 2025-02-12

   $300M Series D at $3B; $50M+ ARR

   235 customers in 42 countries; 40 customers a year earlier

   [5]

2. 2025-06-23

   $300M Series E at $5B

   "our $300M Series E round at a $5B valuation"

   [2]

3. 2025-08-04

   $100M ARR; 42% of the AmLaw 100

   "more than 500 customers in 54 countries"; weekly active users 4x, monthly queries 5.5x

   [3]

4. 2025-11-14

   700 customers in 63 countries; enterprise share from 4% to 33%

   [4]

5. 2026-09-09

   $550M at $15.5B; 80% of the AmLaw 100

   "80% of Am Law 100 law firms use Harvey"

   [1]

6. 2026-09

   About $300M ARR, 960 employees (podcast and press; secondary)

   [10]

Acquisition channels

Five channels: the large-firm referral chain (AmLaw 100 share 42% in August 2025 [3], 80% in September 2026 [1]); the OpenAI Startup Fund seed and early model access [4]; the LexisNexis content partnership [3]; expansion into in-house legal teams, with enterprise share rising from 4% to 33% in a year [4]; Microsoft 365 Copilot, DocuSign and Intapp integrations (Sacra, secondary data) [9]. Five of the Fortune 10 are customers [1].

Loop

A large firm's signature leads the firm's corporate clients to want to work on the same platform; the in-house team becomes a customer and the seat count inside the firm grows. Sacra estimates the median seat count doubles within 12 months [9]. Weinberg treats permissioning as the condition of the loop: "You have to get the permissioning right so agents can access the right systems." [4]

Activation

Entry at roughly $1,200 per lawyer per month with 12-month commitments and about a 20-seat minimum (Sacra estimate, secondary data) [9]. A customer-success staff of former lawyers makes up about a tenth of the team (Sacra, secondary data) [9]; more than 200 lawyers are reported to work in product and go-to-market (secondary data) [10].

Retention evidence

Usage indicators exist: weekly active users 4x, monthly queries 5.5x, active files from 268 thousand to 9.75 million (August 2025) [3]. Net revenue retention and churn are undisclosed. Seat expansion is only a Sacra estimate [9].

Role of pricing

High per-seat price with annual commitment; no outcome-based pricing. (opinion) ARR growth comes mostly from seat expansion; the risk of nominal inflation is lower than in credit models.

Distribution moves

Data coverage for eight new markets and the LexisNexis collaboration in August 2025 [3]; in 2026 an open-weight model and the LAB benchmark under the "own their intelligence" positioning [1].

Team

350 employees (August 2025) [6], about 400 (November 2025) [4], 960 (September 2026, secondary) [10]. (opinion) Roughly $0.3M of annual revenue per employee, a third of ElevenLabs. The human intensity of vertical enterprise sales shows in the number; the ratio is computed from two sources.

The five questions of the Growth Claim Check, applied to this company's public claims.

Test your own claim with the same five questions: Growth Claim Check →

• The valuation is about fifty times the secondary ARR figure: $15.5B [1] against roughly $300M ARR [10]. (opinion) The price buys the continuation of growth in advance.
• A 960-person team with more than 200 lawyers (secondary data) [10]: whether this is a software margin or a services margin is open.
• Model dependency began with OpenAI; the 2026 open-weight model is a move to reduce it [1].
• The price level Sacra estimates is closed to small firms; the market is limited to the upper segment (secondary data) [9].

This section is opinion; a reading, not a source.

1. 01

   (opinion) In a vertical, the referral chain is the channel. 80% of the AmLaw 100; in Turkey the top twenty law firms form a similar chain. Read alongside the Apilex case.

2. 02

   (opinion) Run the quality test early and measurably. "Out of a hundred samples, how many would lawyers send without edits" is a threshold any vertical AI product can copy.

3. 03

   (opinion) Buyer and the buyer's client at the same table. The firm and in-house loop; in Turkey, accountant and SME, insurance agent and client pairs can be built on the same logic.

4. 04

   (opinion) Human-intensive growth eats margin. The ratio may look better with lira costs, but the forward-deployed model carries a scaling limit.

5. 05

   (opinion) Plan to own your model. Its own model after three years; at the start an API, with the moat in workflow and permission architecture. Test question: if your model provider ships the same feature tomorrow, why does the firm stay?

A question for you: remove this company's logo and drop the loop onto your own product. Which node stays empty?

Numbers are the references in the text. An 'opened' source was read by the researcher; 'secondary' ones are verified against their primary source before publication.

1. [1]
   Harvey Raises $550M at a $15.5B Valuation to Help Legal Teams Own Their Intelligence

   Harvey blog · 2026-09-09 · opened

   “80% of Am Law 100 law firms use Harvey”

2. [2]
   Harvey Raises Series E

   Harvey blog · 2025-06-23 · opened

   “Today, we are thrilled to announce our $300M Series E round at a $5B valuation.”

3. [3]
   Harvey Reaches $100m ARR + 42% of AmLaw 100

   Artificial Lawyer · 2025-08-04 · opened

   “more than 500 customers in 54 countries”

4. [4]
   Inside Harvey: how a first-year legal associate built one of Silicon Valley's hottest startups

   TechCrunch (Connie Loizos) · 2025-11-14 · opened

   “On 86 of the 100 samples, two out of three attorneys or more said they would send it with zero edits”

5. [5]
   Legal AI startup Harvey raises $300 million Series D at $3 billion valuation

   Fortune · 2025-02-12 · opened

   “Legal work is incredibly complex and requires so much context, that a simple chatbot doesn't work.”

6. [6]
   Harvey reaches $100m ARR

   Non-Billable · 2025-08-04 · opened

   Note: 350 employees; over $700M total funding reported

7. [7]
   Harvey at $75M ARR

   Sacra · 2025 · secondary

   Note: could not be opened

8. [8]
   Legal AI startup Harvey raises $200 million at $11 billion valuation

   CNBC · 2026-03-25 · secondary

   Note: 403; headline only

9. [9]
   Harvey profile (estimates and round table)

   Sacra · 2026 · opened

   “median seat count doubles within 12 months · $1,200 per lawyer per month with 12-month commitments and roughly 20-seat minimums · $5M seed round led by OpenAI Startup Fund in November 2022 · $21M Series A in April 2023”

10. [10]
    Inside Harvey AI: $11B, $300M ARR, 960 Employees, 12 Offices

    Sourcery (podcast and news) · 2026-09 · secondary

    Note: headline; secondary

• The URL of Harvey's own $100M ARR post; the figure arrives via Artificial Lawyer and Non-Billable.
• An official source for the March 2026 $200M round at $11B (CNBC returned 403).
• Is the September 2026 "$300M ARR, 960 employees, 2,000 customers" official or a podcast claim?
• Pricing ($1,200 per lawyer) is entirely a Sacra estimate.
• Net revenue retention, gross margin, pilot share: none.