Harvey Legal AI Hits $15.5B on $550M Series F Round

Harvey, the legal-tech AI startup, raised
$550 million at a $15.5 billion valuation in a round co-led
by Diffusion and Lightspeed Venture Partners, according to
reporting on September 9, 2026. The round lifts Harvey’s paper value by
roughly 41% from the $11 billion mark it hit in March 2026 and nearly
doubles its December 2025 valuation of $8 billion
(TechCrunch, Sept 9, 2026;
PYMNTS, Sept 9, 2026).

Harvey has now raised more than $1.55 billion in total
across at least eight priced rounds since 2023, one of the fastest
capital-formation arcs in the current AI cycle. The Series F closes just
days after the company introduced Harvey Tenet, its first proprietary
model post-trained on legal data.

Deal terms and lead investors

The round was co-led by Diffusion and Lightspeed Venture Partners. Both
firms were existing believers in the legal AI category; Lightspeed
previously backed Harvey and other application-layer AI companies, while
Diffusion has emerged as an anchor in AI mega-rounds. The company did not
officially label the round as a Series F, though its position in the
capitalization stack and the size of the check make that designation the
consensus read.

Notably, the raise is a primary capital round rather than a
secondary tender offering employees liquidity — meaning the fresh dollars
land on Harvey’s balance sheet for hiring, model training compute, and
international expansion. That distinction matters for valuation
interpretation: primary raises re-price the company for new economic
capital, while secondaries can be one-off tape prints.

Valuation trajectory: from $8B to $15.5B in nine months

The stair-step is aggressive even by 2026 standards. In nine months
Harvey has taken its equity value from $8 billion to $15.5 billion — a
94% mark-up. For context, that pace of paper value creation on primary
rounds sits in the same neighborhood as Anthropic, OpenAI, and xAI over
the same period, and it comes without the training-run capital
requirements that make those foundation-model rounds so large in
absolute dollars.

Traction: 75+ AmLaw 100 firms, 200K+ users

Harvey has crossed a set of adoption milestones that materially change
the story from “interesting legal AI experiment” to “standard
infrastructure for elite law firms.” The company reports it now serves
2,400+ legal
organizations including 75+ of the AmLaw 100 (the top-100 US law firms
by revenue), 200,000+ professional users, and clients in 70+ countries.
Named customers on Harvey’s site include A&O Shearman, Latham &
Watkins, Norton Rose Fulbright, Reed Smith and in-house legal teams at
Deutsche Telekom, Bayer, KKR, Booking.com, Procter & Gamble,
Comcast and Merck.

The product surface has expanded from a research assistant into a
full stack: Vault for document analysis,
Knowledge for research across legal and tax domains,
Agents for end-to-end task execution,
Contract Intelligence for review and negotiation,
plus a Command Center for adoption analytics inside
firms. That last product tells you where Harvey sees the moat: not
just a better retrieval system, but the workflow layer that a CIO buys
firm-wide and instruments.

How this round sits in the AI mega-round tape

Two things stand out on the tape. First, Harvey’s revenue multiple
per dollar of capital raised is far tighter than the pure-compute
players like Crusoe, because it doesn’t need multi-hundred-million-dollar
data-center commitments to grow. Second, its valuation-to-round ratio
(28× on this raise) is closer to a classic Series B/C repriceing than
to the double-digit dilution investors accept for foundation-model rounds
— a sign that primary investors are still comfortable underwriting
software-margin economics for the workflow layer of legal AI.

What could break the story

Three risks are worth flagging without prescribing a view:

• Foundation-model owners moving up the stack. If
  OpenAI, Anthropic, or Google ship domain-specialized legal agents as a
  first-party offering, Harvey’s differentiation shifts from “we have
  the best model prompt-tuned for law” to “we own the workflow.”
  Harvey Tenet — the in-house model reported by TechCrunch — is a hedge
  against that scenario.
• Enterprise legal budgets. AmLaw 100 firms are
  price-insensitive, but the mid-market and in-house corporate legal
  teams that make up the volume tail are more cyclical.
• Model liability. Legal work is a domain where
  hallucination is a fireable offense, and courts have already sanctioned
  attorneys who filed AI-generated briefs with fabricated citations
  (SEC EDGAR filings from public law-firm-adjacent
  filers periodically discuss model-use policies).

What to watch next

The Series F re-frames the near-term calendar for Harvey. Watch for
(i) confirmation of ARR — the last widely-cited figure was in the
low nine-figures; (ii) international expansion beyond the 70+ countries
Harvey lists, especially in the UK and EU where Magic Circle adoption is
already high; and (iii) whether Harvey follows Cognition into the
“quiet IPO” discussion. Late-stage rounds this large usually
buy 18–24 months of runway; the next data point is likely a
customer or product milestone rather than another mark-up.

Sources

• TechCrunch — Harvey hits $15.5B valuation, months after reaching $11B (Sept 9, 2026)
• PYMNTS Investment Tracker — Harvey Raises $550 Million to Bring AI to Law Firms (Sept 9, 2026)
• Harvey company site — product suite, customers, user metrics

Disclosure: This article is for informational purposes only and is not investment advice.