Paris-based Arlequin AI has raised €28 million to develop an alternative neural-network architecture, giving European investors another route into artificial intelligence beyond the increasingly expensive race to build ever-larger foundation models.
Tech.eu reported that Arlequin secured a €28 million Series A round co-led by redalpine and OTB Ventures, with participation from Bpifrance’s Defence Innovation Fund. Existing investors Vsquared Ventures and 10x Founders increased their stakes, while Xavier Niel joined the financing.
Arlequin said the round is backed exclusively by European investors and will fund development of proprietary models based on topological neural networks, or TNNs, alongside the international expansion of its technology.
Arlequin is pursuing a different AI architecture
Founded in 2024 by Hugo Micheron and Antoine Jardin, Arlequin develops AI systems intended for organizations working with large and fragmented datasets.
Tech.eu describes Micheron as a specialist in terrorism studies and geopolitical instability, while Jardin previously worked as a CNRS research engineer specializing in data science and human behavior. Their company’s platform examines relationships across documents, transactions, video and operational data while allowing users to trace findings back to underlying evidence.
The startup is now developing proprietary topological neural networks designed to learn from both individual data points and the relationships connecting them.
Jardin told Tech.eu that Arlequin believes further AI progress will require different architectures rather than simply larger models trained on more data and computing power. The company also says its approach is intended to require less compute than large-scale AI systems.
That makes the technical thesis particularly relevant in Europe, where startups generally cannot match the capital and computing resources available to the largest U.S. frontier laboratories.
Governments are already using the technology
Arlequin sees applications across security and defense, criminal investigations, fraud and money-laundering detection, information integrity, cybersecurity and AI safety.
According to the company, its technology is already used by governments and large organizations across Western and Eastern Europe.
The new capital will support the international team developing and training its models. Tech.eu reported that Arlequin has opened offices in London and Berlin and plans to establish an AI laboratory in Silicon Valley.
CEO Micheron said the company is targeting AI systems capable of understanding complex dynamics hidden across millions of data points.
Europe needs more than another large language model
The significance of the round is not simply that another French AI company raised €28 million.
Much of the frontier-model market has been shaped by a scaling strategy: more training data, more accelerators and larger computing clusters. Arlequin is betting that some high-value problems may instead reward architectures designed around relationships within complex data.
That thesis still has to prove itself commercially and technically against rapidly improving general-purpose models.
But for European founders and investors, Arlequin offers a more interesting signal than the funding amount itself: Europe’s AI opportunity may not depend entirely on reproducing the largest American models. Architectural specialization could become another way to compete where compute alone is difficult to match.
