Wrapper companies were thought to not have very sustainable business models in the early days of the AI revolution, but they now seem to be thriving and turning into serious businesses.
Harvey AI, the legal-focused generative AI platform, announced today a $160 million investment at an $8 billion valuation led by Andreessen Horowitz. The round represents a remarkable leap from the company’s $3 billion valuation during its Series D in February 2025, underscoring growing confidence in vertical AI applications despite early skepticism about companies built on top of foundation models.
The investment follows EQT’s recent backing and includes new investors WndrCo and accounts advised by T. Rowe Price Associates, Inc., alongside returning investors Sequoia Capital, Kleiner Perkins, Conviction, and Elad Gil. As part of the financing, Harvey will complete its first tender offer, providing liquidity to early employees and investors.
Rapid Enterprise Adoption Drives Valuation
The San Francisco-based startup’s momentum is reflected in its client roster: over 50% of the AmLaw 100 law firms now use Harvey, alongside prominent in-house legal teams at Bridgewater, Comcast, and Carvana. The company also counts Fortune 500 clients including AT&T, Verizon, and KKR, with operations spanning 58 countries.
Founded by Winston Weinberg, a former lawyer, and Gabe Pereyra, an AI researcher with stints at Google and Meta, Harvey launched with a clear thesis: that generative AI could be most valuable when trained specifically for complex professional workflows rather than general consumer use. The company builds on custom-trained models from OpenAI, fine-tuned on extensive legal datasets including case law, contracts, and regulatory documents.
Purpose-Built Tools for Legal Work
Harvey’s platform centers on three core products designed for legal professionals. Its AI Assistant provides cited answers to complex legal queries in plain English, grounding responses in authoritative legal sources to ensure accuracy. Vault enables secure analysis of large document sets, critical for due diligence in mergers or compliance reviews, with integration directly into Microsoft Word. The Workflows product automates repetitive tasks such as clause extraction and risk scoring using modular AI models.
The company has also announced plans to expand beyond its OpenAI foundation, with support coming for Anthropic’s Claude and Google’s Gemini models, signaling a multi-model strategy that could help insulate the business from dependence on any single AI provider.
The Wrapper Debate Revisited
Harvey’s valuation trajectory offers a data point in the ongoing debate about “wrapper” companies that build applications on top of large language models. Early critics argued these businesses lacked defensible moats, vulnerable to both foundation model providers moving up the stack and to commoditization as AI capabilities became widely accessible.
Yet Harvey’s growth suggests that deep domain expertise, proprietary training data, enterprise security infrastructure, and workflow integration create substantial value beyond simply packaging API calls. Other wrapper companies, such as Perplexity ($20 billion) and Cursor ($29 billion) have also become extremely valuable. It remains to be seen if these companies can maintain or grow these valuations, but at the moment, there seems to be no stopping them.