Arlequin AI

AI Infrastructure · Defense & Security

Funding

Arlequin AI raises $30M Series A for topological neural networks

The approach traces every conclusion back to source data rather than predicting likely text, targeting audits and evidence LLMs struggle to provide.

• Arlequin AI raised €28 million (about $30 million) in a Series A round co-led by redalpine and OTB Ventures, with Bpifrance's Defence Innovation Fund also participating.
• Founded in Paris in 2024 by former jihadism researcher Hugo Micheron and ex-CNRS data scientist Antoine Jardin, the startup builds AI for governments and large organizations making high-stakes decisions from massive, fragmented datasets.
• Its topological neural networks learn from the structure and relationships within data rather than predicting sequences of words, aiming to capture complex multi-way interactions and trace results back to underlying evidence.
• The platform already serves more than 30 clients across four European countries, including governments and large organizations in Western and Eastern Europe.
• Funds will expand Arlequin's international engineering team, support commercial deployment across Europe, and back new offices in London and Berlin plus a planned Silicon Valley AI lab.
• The bet on a non-LLM architecture reflects Europe's push to build sovereign AI capable of auditable, evidence-based decisions for defence and security, distinct from the transformer models dominating the US and Chinese AI race.