Machenta, an Imperial College London spinout that teaches robots new jobs from a single demonstration, has come out of stealth with $40m in seed funding co-led by Creandum and Union Square Ventures, a round the company says is the largest ever raised by a UK robotics business at seed.

The company was founded in 2025 by Edward Johns, professor of robot learning in Imperial's Department of Computing, and Vitalis Vosylius, a PhD student in the lab Johns runs. The research behind it started about six years ago in Imperial's Robot Learning Lab.

Eleven people work there. They are moving into a new office and lab in King's Cross.

How the model learns from one demonstration

Most robot learning is expensive because it runs on real-world data. A human walks the machine through a task, again and again, and the robot learns that one task. Change the object or the bench layout and much of the work starts over.

Machenta trains its model entirely on simulated data instead, with no real-world human demonstrations in the pre-training at all. The aim is not to teach the robot any particular task but to teach it how to learn, so that one demonstration is enough to pick up something new.

Johns calls the result "physical vibe coding". The company has shown a single model handling packing, tidying and assembly work across different robots, humanoids among them.

"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.

The investors are backing that framing rather than a product. "Robots become so much more useful when they can instantly pick up a new task and get to work right away," said Kathi Wilhelm, partner at Creandum. Rebecca Kaden, managing partner at Union Square Ventures, said the field needs "general-purpose AI which can adapt quickly in the same way that humans can".

A seed round bigger than UK robotics used to raise in a year

The size of the cheque is easier to read against the sector it landed in.

UK robotics companies raised US$258m in 2026 up to late August, across 12 rounds, according to Tracxn data reported in August. In 2024 the whole sector raised US$51m across 19 rounds.

Machenta's seed alone is close to 80% of that 2024 total. On the August tally it would rank third among UK robotics rounds this year, behind Humanoid's US$152m Series A and Automata's US$45m Series C, both of which are later-stage money.

It also moves weight toward London. Cambridge has long led UK robotics, with US$1.3bn raised cumulatively against London's US$582m, even though London has far more robotics companies. A King's Cross lab funded at this level narrows that gap.

As an Imperial spinout, Machenta is also a live example of a route more UK researchers are taking. Universities now take an average stake of about 16% in their spinouts, down from 22% to 25% before the 2023 reforms, as Bizztor's guide to UK university spin-out deals sets out. Neither Imperial nor the company has said what stake the university holds here.

Professor Mary Ryan, Imperial's Vice-Provost for Research and Enterprise, said the round "signals an opportunity for the UK to become a global leader in AI-enabled robotics".

What Machenta is not doing yet

The company is clear that it is a research business and has no plans to commercialise soon. There are no customers or pilots to point to, and the money is going on compute, simulated data and hiring rather than a go-to-market push.

The technical limits are real too. Precise tasks and complex motions remain very hard, which is the gap between a convincing demonstration of tidying and a robot that earns its place on a production line.

That makes this a bet on a research direction, which is unusual at this size in the UK. It sits alongside other deep tech rounds that priced the science rather than the revenue, from Salience Labs' $125m for photonic AI chips to the British Business Bank's £180m into AI and deeptech funds, and adds to London's share of UK startup funding.

For founders watching, the interesting part is not the robot. It is that a simulation-first claim, with no revenue and no product, cleared $40m at seed in London.