Grubel raises €3M Pre-Seed to automate AI specialisation for complex legal matters

Grubel (Munich) has raised €3 million in a pre-seed round led by Point Nine, with a group of named angel investors that includes Jeff Dean (Google's former chief scientist), Chris Ré (Stanford professor and Together AI co-founder), Ion Stoica (UC Berkeley professor and Databricks co-founder), and Gabe Pereyra and Winston Weinberg (both co-founders of Harvey, the legal AI company). Founded in 2026 by Moritz Hardt — director of the Max Planck Institute for Intelligent Systems and former professor at UC Berkeley — and Reinhard Heckel — professor of machine learning at the Technical University of Munich and formerly at IBM Research — the company is building AI systems that specialise themselves to individual legal matters rather than applying a fixed model across all cases.

The specialisation loop

The architectural argument is this: complex legal work — a merger review, a large arbitration, a cross-border regulatory investigation — does not arrive with ready-made training data, task environments, or evaluation criteria that allow a general model to perform reliably. Each matter is structurally novel. Grubel is building an automated loop that handles three sequential problems: discovering and curating the data relevant to the specific matter, adapting the model to that matter's particular requirements, and continuously evaluating outputs against the case's own standards of correctness.

Reinhard Heckel, CEO: "Complex legal matters do not come with ready-made training data, task environments, or clear tests of success. We are automating the AI specialisation loop that identifies the right information, curates it into training data and environments, adapts the system to the matter, and tests whether its work meets the required standard."

The long-term ambition extends beyond law — the co-founders describe complex legal matters as the initial proving ground for an approach they expect to apply to other forms of complex knowledge work. The product stage is not yet publicly disclosed; the research is.

The angel map

The angel roster is the round's credibility signal. Jeff Dean shaped the infrastructure that makes modern large language models possible; Ion Stoica co-built the distributed systems that make them deployable at scale; Chris Ré has done foundational work on data-centric AI. Together they represent three angles of the technical critique Grubel's architecture must survive. Gabe Pereyra and Winston Weinberg, as Harvey co-founders, bring direct commercial experience of what works and does not work when selling AI to legal firms — the hardest part of the Grubel roadmap.

Louis Coppey of Point Nine: "Over the past two to three years, we've seen many legal AI companies emerge, but few are pushing the boundaries of AI research in the legal field as ambitiously as Reinhard, Moritz and the Grubel team. We believe their approach can unlock new capabilities in legal AI and, over time, other forms of complex knowledge work." Point Nine previously backed Aardaia — the Paris-based crop science AI company — in a deal ProYarn covered earlier this year; the common thread is domain-specific AI where general-purpose models encounter structural data gaps that cannot be closed by scaling.