Paris start-up Bleu Robotics raises €5M to train factory humanoids in under an hour

30/10/2026
32
Abstract illustration of stylised robotic arms tracing motion paths over a factory grid

Founded by two INRIA research directors and a former Renault production manager, the Paris start-up says its software can teach a humanoid robot a factory task from about ten demonstrations.

Image: illustrative, AI-generated (no press image supplied).

Paris-based Bleu Robotics has raised €5 million in a pre-seed round to make humanoid robots far quicker to deploy on factory floors. The round was co-led by Contrarian Ventures and OVNI Capital, with participation from Kvanted, WEPA Ventures, Motier Ventures and technology investor Charlie Songhurst, a Meta board member.

The company's stated goal is to cut the full process of deploying a humanoid robot on a new task to one hour or less. Bleu says it will use the funding to grow its team and bring its software to more industries.

Learning from a handful of demonstrations

Most humanoid robots today rely on end-to-end AI models that learn everything inside a single "black box" and need large volumes of task-specific data, often hundreds of hours, before they perform reliably. Bleu takes a different route. Its abductive learning approach infers the intended outcome and constraints of a task from as few as ten operator demonstrations, and the company says it uses up to 100 times less data than alternative methods while keeping the explainability needed for industrial safety and the accuracy above 99 per cent that production lines demand. These are company figures and have not been independently verified.

The software is designed to run on any humanoid platform. Operators teach the robot at the workstation using handheld grippers, without writing code, and can retrain it when production needs change or the robot moves to another site.

"Humanoid robots in manufacturing are the next automotive sector, with the potential to help Europe and the US reshore production," said co-founder Benoît Berkoukchi, who previously managed large assembly-line teams at Renault. "But most robotics companies are trying to train humanoids with millions of data points and build fully general-purpose models. That approach isn't suited to manufacturing. There isn't enough data on factory work, and there won't be internet-scale data, the kind LLMs were built on, for robots for a long time. Our methodology means we can do more with less."

Berkoukchi founded the company with Jean-Baptiste Mouret and Serena Ivaldi, both research directors at France's national computer science institute INRIA. Bleu says initial deployments are under way with major automotive manufacturers and with WEPA, a European hygiene paper producer, and that it is developing partnerships with robot distributors.

Thomas Poitrineau, investor at Contrarian Ventures, said the firm was drawn by the team's decision "to rethink how humanoid robotics software should work from first principles", adding that the approach can put "robots to work in factories within months rather than years".

Why it matters

Humanoid hardware has advanced quickly, but software has lagged. Analyst firm Interact Analysis found that only about 10 per cent of the more than 20,000 humanoids produced last year reached real-world applications. Europe's manufacturers, short of workers and under pressure to bring production home, need robots that can handle high-mix, low-volume work that fixed industrial arms cannot.

Bleu joins a busy European field. EU-Startups counts 2026 rounds for Inbolt, Trener Robotics, THEKER, Humanoid, Sereact, Enact Intelligence and Nomagic among others, totalling about €377.7 million including Bleu's round. Its data-light approach is a clear bet against the general-purpose, data-hungry models favoured by many rivals.

Sources: Bleu Robotics funding announcement, 29 October 2026, as reported by EU-Startups (eu-startups.com), The Engineer and Startup.eu, 29 October 2026; Sifted, October 2026. No embargo applied.

Recent

Cybersecurity News Apps Worth Your Attention

Breaking Tech Europe: Top 5 – Anne Boden Builds a Bank for AI Agents, Whitehall Hands ElevenLabs a £10.8m Voice Deal, and AI Floods Brussels

Microsoft's day-long outage shut Dutch courts' email and Amsterdam's service desks, while Berlin's on-premise servers kept running

AI Meeting Assistants Are Changing Who Gets Heard

Paris start-up Bleu Robotics raises €5M to train factory humanoids in under an hour

30/10/2026
32
Abstract illustration of stylised robotic arms tracing motion paths over a factory grid

Founded by two INRIA research directors and a former Renault production manager, the Paris start-up says its software can teach a humanoid robot a factory task from about ten demonstrations.

Image: illustrative, AI-generated (no press image supplied).

Paris-based Bleu Robotics has raised €5 million in a pre-seed round to make humanoid robots far quicker to deploy on factory floors. The round was co-led by Contrarian Ventures and OVNI Capital, with participation from Kvanted, WEPA Ventures, Motier Ventures and technology investor Charlie Songhurst, a Meta board member.

The company's stated goal is to cut the full process of deploying a humanoid robot on a new task to one hour or less. Bleu says it will use the funding to grow its team and bring its software to more industries.

Learning from a handful of demonstrations

Most humanoid robots today rely on end-to-end AI models that learn everything inside a single "black box" and need large volumes of task-specific data, often hundreds of hours, before they perform reliably. Bleu takes a different route. Its abductive learning approach infers the intended outcome and constraints of a task from as few as ten operator demonstrations, and the company says it uses up to 100 times less data than alternative methods while keeping the explainability needed for industrial safety and the accuracy above 99 per cent that production lines demand. These are company figures and have not been independently verified.

The software is designed to run on any humanoid platform. Operators teach the robot at the workstation using handheld grippers, without writing code, and can retrain it when production needs change or the robot moves to another site.

"Humanoid robots in manufacturing are the next automotive sector, with the potential to help Europe and the US reshore production," said co-founder Benoît Berkoukchi, who previously managed large assembly-line teams at Renault. "But most robotics companies are trying to train humanoids with millions of data points and build fully general-purpose models. That approach isn't suited to manufacturing. There isn't enough data on factory work, and there won't be internet-scale data, the kind LLMs were built on, for robots for a long time. Our methodology means we can do more with less."

Berkoukchi founded the company with Jean-Baptiste Mouret and Serena Ivaldi, both research directors at France's national computer science institute INRIA. Bleu says initial deployments are under way with major automotive manufacturers and with WEPA, a European hygiene paper producer, and that it is developing partnerships with robot distributors.

Thomas Poitrineau, investor at Contrarian Ventures, said the firm was drawn by the team's decision "to rethink how humanoid robotics software should work from first principles", adding that the approach can put "robots to work in factories within months rather than years".

Why it matters

Humanoid hardware has advanced quickly, but software has lagged. Analyst firm Interact Analysis found that only about 10 per cent of the more than 20,000 humanoids produced last year reached real-world applications. Europe's manufacturers, short of workers and under pressure to bring production home, need robots that can handle high-mix, low-volume work that fixed industrial arms cannot.

Bleu joins a busy European field. EU-Startups counts 2026 rounds for Inbolt, Trener Robotics, THEKER, Humanoid, Sereact, Enact Intelligence and Nomagic among others, totalling about €377.7 million including Bleu's round. Its data-light approach is a clear bet against the general-purpose, data-hungry models favoured by many rivals.

Sources: Bleu Robotics funding announcement, 29 October 2026, as reported by EU-Startups (eu-startups.com), The Engineer and Startup.eu, 29 October 2026; Sifted, October 2026. No embargo applied.

Recent

© European Tech On Heels - 2026
Made with
Web Wings