Harvey
Legal AI tailored for leading law firms and corporate legal teams worldwide
Website: https://www.harvey.ai/
Cover Block
PUBLIC
| Field | Value |
|---|---|
| Name | Harvey |
| Tagline | Legal AI tailored for leading law firms and corporate legal teams worldwide |
| Headquarters | San Francisco, CA |
| Founded | 2022 |
| Stage | Growth / Late Stage |
| Business Model | SaaS |
| Industry | Legaltech |
| Technology | AI / Machine Learning |
| Geography | North America (235 clients across 42 countries reported) |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (2): Winston Weinberg, Gabriel Pereyra |
| Funding Label | $100M+ |
| Total Disclosed | Approximately $1B cumulative [TheAIInsider, Mar 2026] |
Links
PUBLIC
- Website: https://www.harvey.ai/
- LinkedIn: https://www.linkedin.com/company/harvey-ai
- Careers: https://www.harvey.ai/company/careers
- Support / Help Center: https://help.harvey.ai/
Executive Summary
PUBLIC
Harvey is a domain-specific AI platform built for law firms and in-house legal teams, and it has become one of the clearest revenue stories in applied generative AI to date. The company was founded in 2022 by Winston Weinberg, a former securities and antitrust litigator at O'Melveny & Myers, and Gabriel Pereyra, a research scientist with prior stints at Google Brain, DeepMind, and Meta [Wikipedia]; [Digidai, Nov 2025]. Its product orchestrates multiple large language models behind a workflow layer designed for contract analysis, due diligence, compliance, and litigation support, and was one of the first investments made by the OpenAI Startup Fund [Harvey Help]; [TechCrunch, Dec 2025]. Annual recurring revenue reportedly crossed $50 million by February 2025 and reached $100 million by August 2025, a roughly six-month doubling that few enterprise software companies of any vintage have matched [Fortune, Feb 2025]; [AllAboutAI]. Capital has scaled in step: a $300 million Sequoia-led Series D at a $3 billion valuation in February 2025, a $160 million round led by Andreessen Horowitz in December 2025 at $8 billion, and reported talks in early 2026 for a further $200 million at $11 billion [Fortune, Feb 2025]; [TechCrunch, Dec 2025]; [Forbes, Feb 2026]. The business model is enterprise SaaS, sold predominantly to large law firms (28% of the Am Law 100 by 2024) and corporate legal departments [Contrary Research]. The next 12 to 18 months will hinge on whether Harvey can defend its early-mover position against legaltech incumbents, sustain net retention as deployments mature, and extend beyond law firms into the broader professional services market it is now publicly targeting [Harvey.ai].
Data Accuracy: GREEN -- Confirmed by TechCrunch, Fortune, Forbes, and Wikipedia.
Taxonomy Snapshot
| Axis | Value |
|---|---|
| Stage | Growth / Late Stage |
| Business Model | SaaS (enterprise seat / subscription) |
| Industry / Vertical | Legaltech, Professional Services |
| Technology Type | AI / Machine Learning (multi-model orchestration) |
| Geography | HQ North America; reported clients across 42 countries |
| Growth Profile | Venture Scale |
| Founding Team | Two co-founders, domain + AI research pairing |
| Funding | $100M+; approximately $1B cumulative reported |
Company Overview
PUBLIC
Harvey began, by the founders' own telling, with a proof of concept in landlord-tenant law and a cold email to Sam Altman [TechCrunch, Dec 2025]. Weinberg and Pereyra were roommates in Los Angeles when they began experimenting with applying large language models to legal work; Weinberg left O'Melveny & Myers after roughly a year of practice to co-found the company in the summer of 2022 [Wikipedia]; [Forbes]. The pitch landed at the right moment. Harvey became one of the first investments of the OpenAI Startup Fund and used that early relationship to ship a product purpose-built for the way lawyers actually work, rather than a generic chat interface bolted onto a foundation model [TechCrunch, Dec 2025].
The company is headquartered in San Francisco and has scaled commercially through enterprise contracts with global law firms. The first reference customer was A&O Shearman (then Allen & Overy), which became the first firm to implement Harvey at enterprise scale in December 2022 [Harvey.ai]. By 2024, third-party tracking placed Harvey inside 28% of the Am Law 100 with 235 client organizations across 42 countries [Contrary Research]. Alan Ghelberg joined as CFO in February 2024, a hire consistent with a company preparing for sustained late-stage fundraising and eventual financial scrutiny [websets.exa.ai].
Milestones since then have come quickly. The Series D in February 2025 valued the company at $3 billion on the back of $50 million ARR; by August 2025 ARR had reportedly reached $100 million; in December 2025 a $160 million round led by Andreessen Horowitz set the valuation at $8 billion; and by February 2026 Forbes reported talks for a further $200 million at $11 billion, subsequently confirmed in March 2026 reporting [Fortune, Feb 2025]; [AllAboutAI]; [TechCrunch, Dec 2025]; [Forbes, Feb 2026]; [SiliconANGLE, Mar 2026]. Cumulative capital raised reportedly reached the $1 billion mark by early 2026 [TheAIInsider, Mar 2026].
Data Accuracy: GREEN -- Confirmed by TechCrunch, Fortune, Forbes, Contrary Research, and Wikipedia.
Product and Technology
MIXED
Harvey markets itself as "the platform built to meet the standards of the world's leading professional service firms," with workflows aimed at contract analysis, due diligence, compliance, and litigation support [Harvey.ai] [PUBLIC]. The product is not a single-model chat assistant. According to Harvey's own help documentation, the system uses a multi-model architecture: "Harvey leverages different advanced AI Large Language Models (LLMs), each designed with unique strengths," and a router "will break down the request into sub-tasks, select a model to use, then synthesize the outputs" [Harvey Help] [PUBLIC]. That orchestration layer, combined with proprietary legal knowledge sources and document storage, is the part of the stack Harvey controls directly.
Distribution and infrastructure run primarily through Microsoft Azure, where Harvey publicly launched its professional services platform in collaboration with Microsoft's Data & AI organization [Harvey.ai] [PUBLIC]. The architecture described in current job listings emphasizes "frontier agentic AI, an enterprise-grade platform, and deep domain expertise," with engineering work focused on "long-horizon (1000+ step) planning agents for mission-critical workflows" (inferred from job postings) [Harvey.ai] [PUBLIC]. The product surface is broadening from a research-and-drafting assistant toward agentic workflows that can carry multi-step legal tasks across systems.
Reported customer outcomes are still early but specific. A&O Shearman has publicly cited staff time savings of two to three hours per week on routine tasks and a roughly 30% reduction in contract review time [Medium] [PUBLIC]. The company also publishes commentary on the changing economics of legal services, including a shift toward "fixed, subscription, retainer, and value-based models emerging alongside traditional billing," framing Harvey as infrastructure for that transition [Harvey.ai] [PUBLIC]. Harvey has not publicly disclosed model-training methodology beyond confirming use of frontier third-party LLMs, and any roadmap items beyond what is published in blog posts and job listings are not addressed here.
Data Accuracy: GREEN -- Confirmed by Harvey.ai, Harvey Help Center, and Medium customer reporting.
Market Research and Opportunity
PUBLIC
Legal services is one of the largest professional services markets in the world, and it is also one of the least automated, which is why generative AI applied at the workflow layer matters now. Global legal services revenue has historically been measured in the high hundreds of billions of dollars annually, and the Am Law 100 alone, Harvey's primary beachhead segment, generates well over $130 billion in annual revenue based on reported industry rankings (analogous market context, industry reporting). A precise third-party TAM specific to legal AI is not present in the verified source set for this report, so we restrict the discussion to demand drivers and adoption signals that are directly cited.
The demand pull is unusually concrete for an AI category. Harvey reports reaching 28% of the Am Law 100 by 2024 and clientele across 42 countries [Contrary Research], and revenue scaled from approximately $50 million ARR in February 2025 to $100 million by August 2025 [Fortune, Feb 2025]; [AllAboutAI]. Customer-cited productivity gains, two to three hours saved per staff member weekly and 30% faster contract review at A&O Shearman, sit at a level of specificity that procurement committees can underwrite [Medium]. Harvey's own commentary points to a structural pricing shift inside law firms toward fixed, subscription, and value-based billing, which historically expands software wallet share when delivery becomes more predictable [Harvey.ai].
Adjacent and substitute markets matter to the thesis. Contract lifecycle management (where Ironclad is a leading independent), document review and e-discovery (Relativity, Everlaw), legal research (Thomson Reuters Westlaw, LexisNexis), and emerging contract-drafting copilots (Spellbook) are all categories where AI capability is being added rather than greenfielded. Harvey's positioning as a horizontal AI platform for legal work, rather than a single-workflow tool, places it in tension with each of these incumbents and with the in-house AI efforts that the largest firms are themselves building. Regulatory and macro forces cut both ways: rising scrutiny of AI use in regulated industries raises the bar for enterprise security and auditability, which favors well-capitalized vendors, while data residency and confidentiality requirements in cross-border legal work create real implementation friction.
| Reported Metric | Value | Source |
|---|---|---|
| ARR (Feb 2025) | $50M+ | [Fortune, Feb 2025] |
| ARR (Aug 2025) | $100M | [AllAboutAI] |
| Am Law 100 penetration (2024) | 28% | [Contrary Research] |
| Client organizations (2024) | 235 across 42 countries | [Contrary Research] |
| Valuation (Feb 2025 → Mar 2026) | $3B → $11B | [Fortune, Feb 2025]; [SiliconANGLE, Mar 2026] |
Analyst takeaway: the table tells a coherent story of revenue and valuation moving together rather than valuation running ahead of fundamentals, which is uncommon among AI companies funded in this cycle. The doubling of ARR in roughly six months is the single most consequential data point in this report.
Data Accuracy: GREEN -- Confirmed by Fortune, Contrary Research, AllAboutAI, and SiliconANGLE.
Competitive Landscape
MIXED
Harvey is positioned as the horizontal AI platform for elite law firms, competing against point-solution legaltech vendors on one flank and against the foundation-model labs and in-house firm builds on the other.
| Company | Positioning | Stage / Funding | Notable Differentiator | Source |
|---|---|---|---|---|
| Harvey | Multi-workflow legal AI platform for Am Law and corporate legal | Late stage; ~$1B raised; ~$11B reported valuation | Multi-model orchestration, A&O Shearman flagship, OpenAI Startup Fund relationship | [TechCrunch, Dec 2025]; [Harvey Help] [PUBLIC] |
| Spellbook | AI contract drafting and review copilot inside Microsoft Word | Venture-backed growth stage | Word-native workflow, mid-market and SMB law firm focus | Structured facts [PUBLIC] |
| Ironclad | Contract lifecycle management with embedded AI | Late stage, well-capitalized | Owns the CLM system of record at large enterprises | Structured facts [PUBLIC] |
The segment map breaks into three layers. At the workflow-software layer, CLM and e-discovery incumbents (Ironclad, Relativity-class players) own systems of record that Harvey does not, and they are racing to add generative capability on top of data they already hold. At the copilot layer, Spellbook and similar drafting tools compete for the everyday document-creation task, often at lower price points and inside Microsoft Word. At the platform layer, Harvey is attempting to be the AI surface across legal work for the largest firms, drawing on its multi-model architecture, its early flagship deployment at A&O Shearman, and its capital base.
Where Harvey's edge looks most defensible today: distribution into the Am Law 100, brand permission with general counsels and managing partners, and the operational know-how of running enterprise-grade legal AI deployments at firms whose security and confidentiality bars are extreme. Capital is also a real moat in this cycle; competing in frontier-model orchestration and enterprise field engineering simultaneously is expensive, and roughly $1 billion raised buys runway few competitors can match [TheAIInsider, Mar 2026]. The perishable parts of that edge are the model layer itself (Harvey does not own the underlying foundation models) and the structural risk that the largest law firms eventually build internal platforms once the workflows are well-understood.
Where Harvey is most exposed: Ironclad sits closer to the contract data than Harvey does in many enterprise accounts, and the foundation-model labs themselves (OpenAI, Anthropic, Google) could push agentic workflow products that compress the value of an orchestration layer. Spellbook owns a channel (Word-native, self-serve, mid-market) that Harvey's enterprise sales motion is not optimized to contest.
The most plausible 18-month scenario: Harvey is the winner if it converts its Am Law penetration into multi-product attach (research + drafting + diligence + agentic workflows) and posts net retention well above 120%; the loser case is one in which a foundation-model lab ships a credible legal agent natively and the largest firms decide a thinner orchestration layer is sufficient, compressing Harvey's pricing power before the platform attach completes.
Opportunity
PUBLIC
If Harvey executes, the prize is becoming the default AI layer for the global legal profession and, by extension, for the broader professional services economy.
The headline opportunity. Harvey's most ambitious plausible outcome is to become the system of intelligence for legal work the way Bloomberg became the system of intelligence for finance: an enterprise tool so embedded in daily workflow that competing against it requires displacing not just software but habit. The cited evidence makes that outcome reachable rather than aspirational. The company has crossed $100 million ARR within roughly three years of founding [AllAboutAI], reached 28% of the Am Law 100 [Contrary Research], and built a reference deployment at A&O Shearman with measurable productivity outcomes [Medium]. Few enterprise software companies in any era have combined adoption breadth, revenue velocity, and brand permission inside the buying segment that matters most to the category.
Growth scenarios.
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Am Law saturation and multi-product attach | Harvey moves from 28% to majority Am Law 100 penetration and lands 3+ products per firm | Continued ARR doubling and contract renewals at expanded seat counts | ARR scaled from ~$50M to $100M in ~6 months [Fortune, Feb 2025]; [AllAboutAI] |
| Corporate legal department expansion | In-house legal teams at the Fortune 500 standardize on Harvey alongside or in place of CLM tools | Microsoft Azure distribution partnership and enterprise security posture | Azure platform launch publicly announced [Harvey.ai] |
| Cross-vertical professional services platform | Harvey extends to accounting, tax, and consulting workflows on the same multi-model architecture | Productization of agentic workflows already described in engineering job listings | Public positioning as a "professional services platform" [Harvey.ai] |
What compounding looks like. The flywheel is forming around three reinforcing loops. First, large-firm reference deployments produce measurable productivity outcomes (two to three hours saved per staff member per week at A&O Shearman) that shorten the sales cycle for the next firm [Medium]. Second, the multi-model orchestration layer accumulates legal-workflow data that improves task routing and output quality without Harvey needing to train a foundation model itself [Harvey Help]. Third, capital scale (roughly $1 billion raised) funds field engineering and security investments that smaller competitors cannot match, which in turn deepens enterprise lock-in [TheAIInsider, Mar 2026]. None of these loops is fully mature, but each is observable in the cited evidence.
The size of the win. A useful comparable is the broader category of vertical AI and legal infrastructure. Thomson Reuters, the closest public peer in legal information services, trades at a market capitalization in the high tens of billions of dollars based on public market data, and has historically commanded premium multiples on recurring revenue tied to legal workflow. If Harvey's reported $11 billion valuation [SiliconANGLE, Mar 2026] sits against $100 million ARR [AllAboutAI], the implied revenue multiple is consistent with the highest-growth enterprise SaaS cohort, and a path to $500 million to $1 billion ARR over the next several years would, at peer multiples, support a valuation materially above current levels (scenario, not a forecast). The downside framing of those numbers, and the execution risks that could compress them, are addressed in the private half of this report.
Data Accuracy: GREEN -- Confirmed by Fortune, AllAboutAI, Contrary Research, Harvey.ai, and SiliconANGLE.
Sources
PUBLIC
[Harvey.ai] Harvey | AI platform for legal and professional services | https://www.harvey.ai/
[Harvey.ai] Meet the team behind our legal AI company | https://www.harvey.ai/company
[Harvey Help] What AI Models Does Harvey Use? | https://help.harvey.ai/articles/what-ai-models-does-harvey-use
[Harvey.ai] Harvey Launches AI-Powered Professional Services Platform on Microsoft Azure | https://www.harvey.ai/blog/harvey-launches-ai-powered-professional-services-platform-on-microsoft-azure
[Harvey.ai] Emerging Trends for the Evolving Business of Law | https://www.harvey.ai/blog/emerging-trends-for-the-evolving-business-of-law
[Harvey.ai] Staff Product Manager role | https://www.harvey.ai/company/careers/8d092528-2554-42e4-a68b-67307b48e6aa
[Harvey.ai] Engineering Manager, Product Engineering role | https://www.harvey.ai/company/careers/e2976ecf-f785-4524-8545-bbd519ffdaae
[Wikipedia] Harvey (software) | https://en.wikipedia.org/wiki/Harvey_(software)
[TechCrunch, Dec 2025] Legal AI startup Harvey confirms $8B valuation | https://techcrunch.com/2025/12/04/legal-ai-startup-harvey-confirms-8b-valuation/
[Forbes, Feb 2026] Legal AI Startup Harvey In Talks To Raise $200 Million At $11 Billion Valuation | https://www.forbes.com/sites/iainmartin/2026/02/09/legal-ai-startup-harvey-in-talks-to-raise-200-million-at-11-billion-valuation/
[Forbes] Harvey | Company Overview & News | https://www.forbes.com/companies/harvey/
[Forbes] Winston Weinberg profile | https://www.forbes.com/profile/winston-weinberg/
[Fortune, Feb 2025] Legal AI startup Harvey lands fresh $300 million in Sequoia-led round | https://fortune.com/2025/02/12/legal-ai-startup-harvey-300-million-series-d-funding-3-billion-valuation-sequoia/
[FT] How a former junior lawyer created a $5bn AI legal start-up | https://www.ft.com/content/49d00498-9a15-4d26-b10c-938bd7e893c6
[Crunchbase] Harvey company profile | https://www.crunchbase.com/organization/harvey-128b
[LinkedIn] Harvey company page | https://www.linkedin.com/company/harvey-ai
[Business Insider, Oct 2025] Harvey's CEO on growing his AI startup | https://www.businessinsider.com/harvey-ceo-grow-legal-startup-winston-weinberg-calendar-audit-hiring-2025-10
[SiliconANGLE, Mar 2026] Harvey raises $200M at $11B valuation | https://siliconangle.com/
[AllAboutAI] Harvey ARR milestone reporting | https://www.allaboutai.com/
[Contrary Research] Harvey research memo | https://research.contrary.com/
[TheAIInsider, Mar 2026] Harvey cumulative funding reporting | https://theaiinsider.tech/
[Digidai, Nov 2025] Gabriel Pereyra background | https://digidai.com/
[Medium] A&O Shearman Harvey deployment outcomes | https://medium.com/
[websets.exa.ai] Harvey office and team data | https://websets.exa.ai/websets/directory/harvey-offices
Articles about Harvey
- Harvey Is Selling Every Am Law 100 Partner a Faster Contract Review — The legal AI startup hit $100M ARR in August 2025 and is now in talks at an $11B valuation, with A&O Shearman as its anchor reference.