The archive · AI & Models · Product decision · 2022–2026
LatticeFlow bets the EU AI Act makes compliance testing a must-buy; COMPL-AI leads
ETH Zurich spinout LatticeFlow published the first EU AI Act LLM compliance framework (COMPL-AI), then turned it into an AI risk-management product.
LatticeFlow AI
What the business is
LatticeFlow (Zurich, Switzerland) is a deep-tech AI risk-management company: it maps regulations like the EU AI Act to technical requirements and evaluates models against them, first through the open-source COMPL-AI framework, now through its commercial AI Assessments platform.
Starting capital：$12M Series A in October 2022 co-led by Atlantic Bridge and OpenOcean with FPV Ventures; over $14M total raised by late 2024, plus a ~$3M Swiss Innosuisse grant in November 2024.
How it started
LatticeFlow was founded in 2022 by Petar Tsankov and ETH Zurich professor Martin Vechev, starting from AI data-quality and robustness research. When the EU AI Act passed, the team moved to define what compliance actually means for large language models, collaborating with Bulgaria's INSAIT institute.
What happened
On 2024-10-16 LatticeFlow, ETH Zurich and INSAIT released COMPL-AI: the first technical interpretation of the EU AI Act plus an open-source LLM evaluation framework built on 27 benchmarks. The launch evaluated models from OpenAI, Meta, Google, Anthropic and Alibaba and found most score poorly on fairness and cyberattack resilience; the European Commission welcomed it as a first step. In November 2024 Innosuisse granted about $3M, and in October 2025 LatticeFlow announced an integration with Vanta to connect technical assessments to governance, risk and compliance workflows.
How it ended up
Still running: LatticeFlow continues as an AI governance platform, with COMPL-AI as its open flagship and commercial technical assessments for enterprises.
Background
LatticeFlow is a Zurich startup betting that the EU AI Act turns AI compliance into a measurable, buyable service. Its core claim: the Act states obligations in high-level terms, so the company that first translates them into concrete technical checks gets to define what compliance means in practice.
Founded in 2022 by Petar Tsankov and ETH Zurich professor Martin Vechev, LatticeFlow released COMPL-AI on 2024-10-16 with ETH Zurich and INSAIT: the first technical interpretation of the EU AI Act for generative models, an open-source evaluation framework with 27 benchmarks, and a public leaderboard scoring models from OpenAI, Meta, Google, Anthropic and Alibaba. TechCrunch and SiliconANGLE covered the launch, and the European Commission called it a welcome first step.
The launch found most models score only around 50% on fairness and cyberattack-resilience benchmarks, while copyright and privacy remain hard to evaluate — evidence that compliance, not just capability, was a real gap. LatticeFlow then commercialized the work: about $3M from Innosuisse in November 2024 and an October 2025 integration with Vanta to feed technical evidence into GRC workflows.
As of September 2026 LatticeFlow is still running as an AI governance platform, riding enforcement of the AI Act whose GPAI obligations started applying in August 2025.
What has to be true
- A regulation nobody can technically measure creates a vacuum; LatticeFlow filled it with a concrete 27-benchmark interpretation.
- Open-sourcing COMPL-AI made the ETH Zurich/INSAIT interpretation the reference point, and the Commission's welcome quote gave it official credibility.
- The public leaderboard turned compliance into a comparable metric — exactly the evidence procurement teams and regulators need.
- Research roots plus a Vanta partnership moved the framework from academic tool to enterprise workflow.
What can be applied
When a regulation is high-level, the first credible technical interpretation becomes the benchmark; publish it openly to build the community, then sell the commercial platform.
Aftermath
As of 2026-09-02 LatticeFlow is live and selling AI governance: it combines the COMPL-AI heritage with its commercial AI Assessments platform, integrates with Vanta so vendors can share regulator-grade evidence, and positions itself for the EU AI Act's enforcement deadlines, which began applying to general-purpose AI providers in August 2025. The company remains Zurich-based, backed by more than $14M in venture funding plus Innosuisse support.
Sources
- This LLM framework takes a first stab at benchmarking Big AI's compliance with the EU AI Act
- LatticeFlow releases framework for checking LLMs' compliance with the EU AI Act
- ETH Zurich, INSAIT, and LatticeFlow AI Launch the First EU AI Act Compliance Evaluation Framework for Generative AI
spotted an error? The archive wants to know.
Your turn
You just read one. Describe what you are building, and see who is betting on the same thing.
Free account · 3 free questions · no card