Machenta raises $40M to teach robots tasks from one demo

The founders and the team

Edward Johns, the chief executive, is a professor of robot learning at Imperial and directs its Robot Learning Lab. According to his Imperial profile, he was a founding member of the Dyson Robotics Lab. Vitalis Vosylius, the chief technology officer, was first author of "Instant Policy," an ICLR 2025 paper on learning new tasks from one or two demonstrations, trained on pseudo-demonstrations generated in simulation.

Johns's lab also published "Learning a Thousand Tasks in a Day" in Science Robotics in November 2025, reporting 1,000 tasks taught in under 24 hours.

Machenta's team page lists researchers from DeepMind, Dyson, Huawei and Stability AI. Headcount is small: Tech.eu put it at around a dozen, while The Cap Table reported 11 staff.

"Intelligence is not so much about what you can already do, it's about how quickly you can learn to do something new," Johns said in Imperial's announcement. "We've spent years trying to understand what it would take for robots to learn at the speed of a human."

What investors say, and where the money goes

Rebecca Kaden, managing partner at USV, was quoted by Tech Funding News taking aim at the prevailing approach. "The industry is pouring enormous amounts of capital into collecting real-world data and training robots one task at a time, but Machenta is changing this," she said.

Kathi Wilhelm, a partner at Creandum, wrote on the firm's blog that "very few teams in robotics can point to a body of work like this."

Tech Funding News reported that the money will go toward hiring, computing power and generating more simulated training data. Imperial said Machenta is moving into a new office and lab in King's Cross.

A crowded, expensive field

Machenta's simulation-first approach stands apart from better-funded rivals that lean on real-world data. TechCrunch reported that Skild AI raised a $1.4 billion Series C at a valuation of about $14 billion in January, and that Physical Intelligence was reportedly in talks to raise about $1 billion at a valuation above $11 billion.

Dream Labs, which describes a zero-shot model called Dream-1 (one that attempts tasks without task-specific examples), raised a $90 million seed. MekaWire ranked Machenta's round the fifth-largest robotics seed or pre-seed globally over the past 18 months, behind rounds that include Walden Robotics ($300 million) and General Intuition ($133.7 million).

Citing Crunchbase, Tech Funding News put physical AI funding at about $47.4 billion across 521 deals in the first half of 2026.

What comes next

Machenta said it is hiring across robot learning, control, simulation and 3D transformers. As of October 28, USV had not published a blog post on the deal.