Arlequin AI raises €28M Series A to build a European alternative for complex-data intelligence

Arlequin AI (Paris) has raised €28 million in Series A funding, co-led by Swiss venture firm Redalpine and Warsaw-based OTB Ventures, with participation from Bpifrance's Defence Innovation Fund, existing investors Vsquared Ventures and 10x Founders increasing their stakes, and Xavier Niel joining the cap table. The company, founded in 2024 by Hugo Micheron and Antoine Jardin, is building AI systems designed for organisations that need to make high-stakes decisions from large, fragmented, and heterogeneous datasets.

The round brings Arlequin's total funding to approximately €32.4 million, following a €4.4 million seed in June 2025. At €28 million from a company less than two years old, this is one of the largest Series A rounds raised by a French deeptech startup in 2026.

The architecture

The central claim Arlequin is making is architectural: that the next generation of AI systems for complex analytical work requires a fundamentally different approach, not merely larger models or more compute. Antoine Jardin, CTO and co-founder — a former CNRS research engineer specialising in data science and human behaviour — argues that systems based on topological neural networks can learn not only from individual data points but from the relationships between them, capturing multi-way interactions that standard architectures represent poorly as datasets scale.

In practice, the platform analyses connections across documents, transactions, video, and operational data simultaneously, allowing users to identify non-obvious relationships within a corpus and trace every finding back to the underlying evidence. The human-verifiable provenance is critical for the client profile Arlequin is targeting: large institutions where an answer without an audit trail is not actionable.

Named client sectors include fraud and money-laundering detection, cybersecurity, information integrity, and complex decision intelligence for government and enterprise organisations across Western and Eastern Europe. The company explicitly positions itself as a European alternative to US-built data intelligence platforms — sovereign architecture, European data governance, and reduced compute dependency relative to large-scale language models are the stated differentiators. CEO Hugo Micheron: "Today, another revolution is taking shape: the development of new AI systems capable of understanding highly complex dynamics hidden within millions of data points."

The round and the expansion thesis

Redalpine brings a track record in deep-tech bets with long development timelines; OTB Ventures adds Eastern European distribution and government relationships. The Bpifrance Defence Innovation Fund's presence signals French state interest in sovereign AI infrastructure — the fund backs companies whose technology has clear applications in critical national infrastructure, not exclusively defence hardware. Vsquared Ventures and 10x Founders re-upping is the standard signal of early-investor confidence in trajectory.

The capital will fund three parallel tracks: building the proprietary model stack (currently the team has around eight researchers out of roughly 30 total), commercial deployment with European and international clients, and geographic expansion. Offices in London and Berlin are already open; an AI research laboratory in Silicon Valley is planned in the coming months.

The international footprint is unusual for a company at this stage. Running R&D, sales, and a research lab across four time zones before reaching Series B requires either a very capital-efficient model architecture or a client pipeline large enough to justify the overhead. The company has been deployed by governments and major organisations but has not disclosed revenue figures publicly.