# Paris's Bleu Robotics Raises EUR 5M Pre-Seed to Teach Industrial Humanoids New Tasks From Operator Demonstrations in About 20 Minutes
> Bleu Robotics raised a EUR 5M pre-seed led by Contrarian and OVNI to teach industrial humanoid robots new tasks from demonstrations in about 20 minutes.
- **Author:** Swati k jagtap
- **Published:** 2026-10-29
- **Updated:** 2026-10-29
- **Category:** Startups
- **Editorial type:** news
- **Topics:** humanoid-robots, ovni-capital, pre-seed-funding, industrial-automation, french-startups
- **Canonical URL:** https://www.todaysstartupnews.com/startups/bleu-robotics-raises-5-million-pre-seed-contrarian-ovni-humanoid-robot-learning
## Key takeaways
- Paris-based Bleu Robotics has raised EUR 5 million in pre-seed funding co-led by Contrarian Ventures and OVNI Capital
- Kvanted, WEPA Ventures, Motier Ventures and Charlie Songhurst also took part
- The company builds AI models that let industrial humanoid robots learn new tasks from operator demonstrations without large task-specific datasets
- Bleu says about 20 minutes of training can lead to a first deployment in under an hour
Bleu Robotics, a Paris-based provider of software and AI models for industrial humanoid robots, has raised EUR 5 million in pre-seed funding. Contrarian Ventures and OVNI Capital co-led the round, announced on October 29, 2026. Kvanted, WEPA Ventures, Motier Ventures and angel investor Charlie Songhurst also took part.

The company says it will use the money to grow its team, expand its research and robotics teams and scale its software across industries to deploy its first industrial robot-learning models.

Bleu Robotics was founded by Benoît Berkoukchi, Jean-Baptiste Mouret and Serena Ivaldi. Mouret and Ivaldi are researchers at the French research institute INRIA, and Berkoukchi is an industrial expert. The startup builds on 20 years of research in humanoid robotics and AI.

Its platform uses what it calls abductive learning technology. It lets humanoid robots learn new industrial tasks from operator demonstrations, without needing large task-specific datasets or long training cycles. The software runs on any humanoid platform, alongside specialised training devices.

The company is aiming at factories where production runs are small and varied. In those settings, the cost and complexity of reprogramming robots have kept automation out of reach. Bleu's approach is to shift from programming to learning.

According to the company, about 20 minutes of training can be followed by a first deployment in less than an hour. Bleu's website describes algorithms that enable robots to learn new operations in about one hour. These figures are the company's own. No customer results have been published.

Bleu launched its first model, Bleu_0.1, in March 2026. It uses few-shot imitation learning, in which a robot learns a complex task from a small number of demonstrations. The software is designed to be compatible with many robot models. Reports list a collaboration with Niryo from May 19, 2026, which provides real hardware for training, testing and industrial rollout, and a partnership with the innov8 Group.

Bleu has won the AI Pioneers award under the France 2030 programme and received a 2026 i-Lab grant from Bpifrance. It describes its offer as a sovereign European software solution that reduces dependence on non-European technologies.

The pre-seed capital will support growth across the research and robotics teams and the roll-out of Bleu's first industrial robot-learning models. The company did not disclose its valuation or customer names.