Legal AI Startup Harvey Secures $550M at $15.5B Valuation

Legal artificial intelligence developer Harvey has raised $550 million in a new financing round that values the company at $15.5 billion, as first reported by TechCrunch Startups on Wednesday. The significant capital infusion highlights sustained investor enthusiasm for domain-specific enterprise artificial intelligence software.

The financing was co-led by venture firms Diffusion and Lightspeed Venture Partners. Harvey opted not to formally label the transaction with a Series letter, mirroring recent fundraising structures utilized by other high-valuation technology developers such as Databricks.

The new valuation represents a substantial escalation over a short timeframe, effectively doubling the company's private market worth in approximately nine months. The deal follows a rapid sequence of capital raises completed by the startup over the past year.

Harvey previously closed a $200 million transaction at an $11 billion valuation in March, which had followed an $8 billion valuation established during a December funding event. Data from PitchBook indicates that Harvey has completed at least eight priced investment rounds since 2023, including five since 2025. Given previous extension rounds, industry trackers view this latest transaction as roughly equivalent to a Series F round.

With the additional $550 million in capital, Harvey's cumulative funding has surpassed $1.55 billion. The expanded balance sheet arrives as the company broadens its footprint beyond legal software applications into underlying AI model architecture.

The fundraising follows the launch of Harvey's initial custom model, dubbed Harvey Tenet, announced several weeks ago. The specialized model was developed utilizing the Kimi K3 open-weight base architecture, which was subsequently post-trained on legal data to suit domain-specific workflows.

To build and deploy Harvey Tenet, the company collaborated with inference technology provider Fireworks. Fireworks previously assisted code-generation company Cursor in post-training its own customized in-house model.

In addition to deploying its own native model, Harvey is urging its legal industry clients to implement and post-train custom open-weight models. The operational shift demonstrates how the legal sector can adopt advanced artificial intelligence applications while reducing its reliance on proprietary frontier models from developers like OpenAI and Anthropic.