London robotics startup Machenta came out of stealth on 27 October 2026 and announced a $40M seed round, as reported by Tech.eu. Creandum and Union Square Ventures co-led the round. According to the company, it is the largest seed round a UK robotics company has raised so far.
What Machenta says it is building
The startup is developing an AI model that lets robots pick up new physical tasks quickly. It positions this against current approaches, which it describes as costly and slow because they need many demonstrations. Co-founder Edward Johns describes the approach as "physical vibe coding", meaning the model learns a task straight away.
Machenta says it trained the model entirely in simulation, using computer-generated environments and no real-world data. Once trained, the model is said to be able to direct a robot through varied jobs, including packaging, assembling and tidying, after seeing the task done a single time. These are the company's own claims; the source article does not report independent testing or a customer deployment.
Who is behind it
Johns is a professor of robot learning at Imperial College London. He co-founded Machenta with his PhD student Vitalis Vosylius. The research began six years ago at Imperial, where Johns was asking whether a robot could learn new physical skills as fast as a person.
What the money is for
Machenta plans to use the funding to scale the technology and grow its team, currently around a dozen people, at its King's Cross office in London.
What is still unknown
The announcement gives no product name, release date, pricing, or list of customers or pilots. It also does not say how the model performs outside simulation beyond the company's own description. For now this is a funding event and a set of claims about a technique, not something a business can deploy.