Harvey confirms $11B valuation: Sequoia triples down

Harvey raised $200M at an $11B valuation, bringing total funding to over $1B and driving its valuation up more than 3.5x in the past year. The round was co-led by returning investors GIC and Sequoia, with participation from Andreessen Horowitz, Coatue, Conviction Partners, Elad Gil, Evantic, and Kleiner Perkins; Sequoia has now co-led three rounds since Series A. Prior marks were $8B in December (a16z-led), $5B in June (Kleiner/Coatue), and $3B in February 2025 (Sequoia-led).

Analysis

AI-first entrants into legal workflows are changing the capture points in the legal stack: the biggest second-order winner is compute and inference — firms that sell GPUs, managed model inference, and embeddings services will see disproportionate incremental spend from large language-model (LLM) adoption in highly paid verticals. Expect enterprise cloud line items tied to usage (GPUs, network egress, managed model APIs) to grow faster than classic SaaS subscription lines for customers that pivot to consumption-based pricing; that compresses short-term SaaS gross margins but increases long-run TAM for infra providers.

Incumbent legal information and workflow vendors face a two-front pressure: faster time-to-answer from LLM workflows reduces per-user usage of legacy research products, while startups drive premium pricing for specialized workflows and integrations. Adoption hinges on three catalysts — demonstrable error rates in live matters (malpractice headlines), contractual procurement wins at global law firms, and data/IP litigation outcomes — each capable of swinging sentiment within weeks-to-months and fundamental revenue trajectories over 6–24 months.

The froth in late-stage private valuations creates tactical opportunities and risks: a financing-driven valuation gap raises M&A probability (acquihires or strategic cloud partnerships) but also sets up sharp re-ratings if enterprise ROI or regulatory friction shows up. Monitor leading indicators (net-new legal enterprise deals, cloud spend per customer, and any malpractice or regulatory enforcement headlines) as 30–90 day triggers that will distinguish durable winners from overstretched private comps.

Key Decisions for Investors

• Long NVDA (NVIDIA) — 6–18 month horizon. Buy NVDA equity or 9–12 month call spreads (financing with high-premium selling discouraged). Rationale: capture outsized GPU/inference demand from legal AI; target position size 4–7% portfolio. Risk: model commoditization or cycle-driven GPU oversupply; set tactical stop at 20% drawdown and trim into 30–50% gains.
• Pair trade: Long MSFT (Microsoft) / Short RELX (RELX) — 3–9 month horizon. Long MSFT to capture enterprise cloud + managed AI adoption; short RELX to express risk to legacy legal research pricing. Size net exposure modest (e.g., 3% long MSFT funded by 2% short RELX) for ~1.5–3x expected asymmetric payoff if cloud capture accelerates. Catalyst: quarterly cloud spend prints and Legal division subscription metrics; unwind if RELX reports accelerating enterprise AI monetization.
• Event/merger-arbitrage style: Buy 6–12 month puts on select high-valuation private-market-exposed SaaS peers or rotate into small-cap legal SaaS names that are likely acquisition targets — implement via options or small-cap longs with 6% portfolio allocation. Rationale: protect against a private-valuation compression wave; payoff if M&A fails to materialize and mark-to-market multiples fall. Risk: wave of strategic acquisitions could re-rate targets higher; keep position sizing disciplined.