From ETH Labs to $300M Valuation: How DeepJudge’s $41.2M Round Showcases Swiss AI Excellence

The Zurich-based legal AI startup’s Series A demonstrates how Switzerland’s research institutions, funding mechanisms, and talent density create global category leaders

DeepJudge, the Zurich-based legal AI startup founded by former Google researchers, has closed a $41.2 million Series A round led by Felicis and prior investor Coatue, valuing the company at $300 million. The funding marks a significant milestone for Swiss deep tech and demonstrates how the country’s innovation ecosystem transforms academic research into globally competitive companies.

Founded in 2020 by three ETH Zurich PhD graduates (Paulina Grnarova, Kevin Roth, and Yannic Kilcher), DeepJudge builds customized AI-powered search engines that help law firms unlock decades of proprietary knowledge trapped in their archives. The company now serves elite global firms including Freshfields, Holland & Knight, and Gunderson Dettmer, having expanded from its initial Swiss customer base to capture the competitive international market.

The Problem DeepJudge Solves

While ChatGPT and Google Search excel at parsing public information, law firms face a different challenge: accessing the massive volumes of confidential documents they’ve accumulated over decades. “Currently, every single law firm has access to the same large language models like ChatGPT, to the same public information on the internet, but what the others don’t have is the data that they have produced themselves (that is their true competitive power),” Grnarova told Forbes.

As Grnarova emphasized in DeepJudge’s funding announcement, “The legal AI race is not about who has the flashiest chatbot. It is about unlocking the trillion-dollar knowledge trapped inside law firms.”

DeepJudge creates a customized index of each firm’s data (similar to the index underlying Google Search) that catalogs what words and phrases appear in which documents, categorizes similar types of deals or clients together, and enables natural language queries across disconnected systems. The platform can answer questions like “Have we negotiated this term before?” or “What has the firm done for this client?” by searching across document management systems, billing data, SharePoint, emails, and matter intake simultaneously.

The index also addresses AI hallucinations by providing extensive context to large language models. This is critical: more than 120 cases of hallucinations in court filings have been identified since 2023, according to law firm Baker Donelson. DeepJudge’s “retrieval-first” approach ensures that AI models have accurate, permission-respecting access to firm knowledge before generating responses.

The Swiss Innovation Pathway

DeepJudge’s trajectory from university project to $300 million company exemplifies the strength of Switzerland’s deep tech ecosystem. The company’s success rested on several pillars unique to the Swiss AI landscape:

World-Class Research Foundations

The three founders met while pursuing PhDs in AI at ETH Zurich, one of Europe’s premier technical universities. Their advisor, who had founded a legal tech company, encouraged them to take courses in the legal department. This interdisciplinary approach (combining deep technical expertise with domain-specific knowledge) proved essential. What began as a class project to create customized search engines for law firms quickly attracted interest from Swiss legal practices willing to become early customers.

“Founded by ex-Google search engineers with PhDs in AI from ETH Zurich, our team’s rigorous technical background in information retrieval is tailor-made for this moment,” Grnarova noted in the company’s announcement. This combination of Google-scale engineering experience and Swiss academic rigor created a technical foundation that competitors struggle to replicate.

Strategic Non-Dilutive Funding

Before approaching venture capitalists, DeepJudge secured early-stage support through Switzerland’s philanthropic and government funding mechanisms. The company won all three stages of Venture Kick in 2021, receiving CHF 150,000 to advance its AI-powered legal document processing platform. This non-dilutive capital allowed the founders to validate their technology and business model before raising equity.

Venture Kick represents a distinctive feature of the Swiss ecosystem. These early grants reduce financial pressure on founders and enable them to maintain equity ownership during critical development phases.

DeepJudge also received CHF 150,000 from Innosuisse’s InnoBooster program in June 2021, further demonstrating how government-backed innovation funding complements private investment in Switzerland.

Access to World-Class Talent and Advisors

The founders’ Google experience proved invaluable, but Switzerland’s concentration of AI expertise provided ongoing advantages. The country ranks among the top five globally for AI research according to the Global AI Index, with research projects funded by the Swiss National Science Foundation more than quadrupling between 2017 and 2022.

DeepJudge assembled a team of legal tech veterans including Tony Ensinger (previously Head of Sales at Casetext, acquired by Thomson Reuters) and Steve Obenski (former senior leader at Kira Systems, acquired by Litera). The company’s advisors include Dr. Thomas Hofmann, former Director at Google and Professor of AI at ETH Zurich, and prominent Swiss business leaders like Dr. Felix Ehrat, former General Counsel of Novartis.

This combination of technical depth and industry expertise reflects Switzerland’s unique position: a small country with outsized concentrations of both cutting-edge research and global corporate headquarters.

Proving Ground: The Swiss Market

Switzerland’s sophisticated legal market provided an ideal testing ground. Homburger, one of Switzerland’s largest business law firms, became an early adopter following a successful pilot. The adoption rate proved remarkable: more than 80% of Homburger’s legal professionals incorporated DeepJudge into their workflow, with engagement levels unparalleled compared to other legal tech tools at the firm.

“For some time we’ve been looking for an advanced search engine that can effectively navigate through millions of internal documents in various formats, so that our collective know-how and expertise are at our lawyers’ fingertips,” said David Oser, partner in M&A at Homburger.

Lenz & Staehelin and CMS Switzerland also became early customers, providing the validation and case studies necessary for international expansion. This early customer validation, achieved in Switzerland’s demanding legal market, provided credibility when approaching global firms.

From Seed to Series A: Building Momentum

DeepJudge raised a $10.7 million oversubscribed seed round in June 2024, led by Coatue with participation from notable angel investors including Gokul Rajaram and Michele Catasta. That round enabled U.S. expansion and product development, including the launch of DeepJudge AI Workflows in March 2025.

The company’s recognition within Switzerland provided additional momentum. DeepJudge placed third in the 2025 Top 100 Swiss Startup Awards, behind engineering startup Corintis and cleantech innovator DePoly. The ranking, which has evolved into a cornerstone of Switzerland’s innovation ecosystem over 15 years, provides critical visibility and credibility.

Coatue partner Caryn Marooney, who led both the seed and Series A rounds, emphasized DeepJudge’s technical approach: “When you are searching and indexing your own information, you’re creating all of these layers of context so when it’s being put into AI, it takes the hallucinations down so dramatically.”

Market Traction and Global Recognition

Between the seed and Series A rounds, DeepJudge demonstrated exceptional execution. The company grew revenues by more than 500% over the past year, a growth rate that attracted Felicis to lead the Series A.

The company’s platform achieved rare adoption metrics for enterprise software: 80 to 90% active usage across its customer base. This stands in stark contrast to typical legal tech tools, which often struggle with single-digit adoption rates.

External validation followed. DeepJudge was recognized as the #1 Legal AI tool in a global law firm survey by SKILLS.law and identified as the likely category leader within one to three years by Legal Technology Hub. These recognitions reflect not just technological capability but real-world impact.

The Series A announcement brought a marquee customer win: Freshfields, a Magic Circle global elite law firm that advises the world’s leading multinational corporations and financial institutions, selected DeepJudge as a core component of its AI and knowledge strategy after a detailed evaluation process.

“Innovation at Freshfields is about strengthening how we work and preparing for what’s next. DeepJudge supports that by enabling responsible, targeted use of proprietary knowledge and powering bespoke AI workflows tailored to our standards,” said Gil Perez, Chief Innovation Officer at Freshfields.

DeepJudge’s customer roster now includes Holland & Knight (an Am Law 100 firm with more than 2,200 lawyers across 34 offices worldwide), Cozen O’Connor (an international firm with more than 825 attorneys across 30 offices), ArentFox Schiff (a leading national law and lobbying firm with over 650 lawyers), and Schoenherr (a tech-forward full-service firm with a strong footprint across Central and Eastern Europe).

What Comes Next

With the new capital, DeepJudge plans to expand its engineering team and accelerate international market penetration. “We’re squarely entering this act two about being a bit louder, getting this in the hands of more lawyers and hiring a bigger sales team,” said Felicis partner James Detweiler.

For Switzerland’s innovation ecosystem, the $300 million valuation and exceptional customer traction represent validation that the country’s approach (combining research excellence, strategic funding, and talent density) can produce AI leaders capable of winning against well-funded competitors on technical merit and product execution.

“The future of legal AI is not generic. It is secure, precise, context-rich, and deeply integrated into the knowledge that makes your firm unique,” Grnarova concluded in the company’s announcement. “We are just getting started.”

That ambition, backed by Swiss engineering rigor and a proven go-to-market strategy, positions DeepJudge as a case study in how Europe’s innovation hubs can compete globally in the AI era.