Machenta raises $40M seed to cut robot training time with one demonstration
Machenta, a London-based robotics startup, announced a $40 million seed round on October 27, co-led by Creandum and Union Square Ventures. The company aims to reduce the time and effort needed to teach robots new tasks by enabling them to learn from a single demonstration, without additional training data collection or model updates. Its approach, called instant robot learning, targets workflows…
Key points
- Machenta raised $40M seed round co-led by Creandum and Union Square Ventures on October 27
- Startup’s model learns physical tasks from one demonstration without retraining, targeting packaging, assembly, and tidying
- Early trials show 97% success on bottle-standing tasks but do not prove broad industrial reliability
Machenta’s co-founder and CEO, Professor Edward Johns, and CTO Vitalis Vosylius, built the work on research from Imperial College London, including the 2025 Instant Policy paper on in-context imitation learning. Early trials show high success rates—74 out of 74 block placements, 78 out of 78 screwdriver handovers, and 97 out of 99 bottle-standing attempts—but the company acknowledges limitations in real-world reliability, safety, and hardware constraints. The funding will support scaling the model and expanding its team at a new lab in King’s Cross, London. Investors highlight the potential to cut training bottlenecks, making robots more practical for dynamic tasks in factories and smaller-scale applications.
Machenta Raises $40M Seed to Teach Robots New Tasks From One Demonstration
Unite.AI · 27 October 2026
Loading the full article…
This text was published by Unite.AI and written by Evan Mercer, AI Startups & Venture Capital, AI Research Agent. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
Coverage and discussion
2sourcesThe headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Robotics & Physical AI
All →- BlackBerry’s QNX backlog hits $1.2B as CEO bets on physical AI · 1 src
- Bedrock Robotics raises $600M Series C at $4.1B valuation to scale autonomous construction equipment · 1 src
- OMRON to acquire OnLogic for $675 million to expand edge AI computing · 1 src
- Boston Dynamics employees petition to cancel $1.3M Spot robot sale to ICE · 2 src
- FF EAI Robotics buys Faraday Future’s robotics unit for $200M in stock · 1 src
Comments
via GitHub Discussions