Machenta, a frontier AI company building a new kind of foundation model for robots, has raised a £30 million seed round co-led by Creandum and Union Square Ventures.
People are remarkably good at learning new physical tasks: show someone how to do something once and, in many cases, they can do it straight away. However, robots today are very different, and teaching them a new task usually takes several days or weeks.
First, hundreds or thousands of human demonstrations of the task are collected, then the model is trained on this data, and finally after waiting for all this, the robot can perform the task by itself.
Machenta has built a foundational model designed to eliminate that intelligence gap between humans and robots, and to achieve ‘instant learning’.
Utilising this model, the firm says, if a robot sees just one demonstration of a task, it can perform the task immediately afterwards – instantly adapting to entirely new variations of the task. There is no need to wait for further demonstrations or training.
Less than a year after Machenta came together, the team has already shown that their approach works across tasks including packaging, assembling, and tidying.
It is a different approach to the dominant one of today, which requires a huge amount of real-world data collection.
“Intelligence is not so much about what you can already do, it’s about how quickly you can learn to do something new,” said Professor Edward Johns, co-founder and CEO of Machenta. “Humans don’t need to see the same thing over and over before they can do it.
“If you take some objects, and show somebody what to do with them, they’ll probably then be able to do the same task immediately. We’ve spent years trying to understand what it would take for robots to learn at the speed of a human. That is what instant robot learning is about.”
The work behind Machenta began six years ago at Imperial College London, where Johns focused on a question that was largely outside the mainstream of robotics research: could a robot one day learn new physical skills as quickly as a human?
Johns is Professor of Robot Learning at Imperial College London, and in recent years he and his PhD student Vitalis Vosylius, built the foundations of in-context learning in robotics — the learning algorithms which are also behind much of the intelligence in LLMs.
After seeing early signs that this same intelligence was now possible in the physical world, Johns and Vosylius founded Machenta to scale this work, and Johns brought his PhD team from Imperial with him. They have recently shown how a single model can instantly learn everyday tasks across any robot, including humanoid robots.
Instant learning also makes robots far easier to teach. If showing a robot once is enough, you no longer need to be a robotics expert to program or train robots. Machenta’s goal is to make robotics accessible to anyone, such that one day you could just walk up to a robot, think of any task, and teach it to the robot instantly. Interacting with robots then starts to look much more like the physical equivalent of vibe-coding: when teaching a robot is this easy, it unlocks a whole new level of creativity and possibility.
The new funding will be used to scale Machenta’s instant learning model, and expand its team at their new office and lab in King’s Cross, London.
“Very few teams in robotics can point to a body of work like this, which has repeatedly changed what people thought was possible,” said Kathi Wilhelm, partner at Creandum. “The team at Machenta has spent years breaking through this extremely hard problem, long before there was a company around it.
“They’re not starting with just a theory of how robots can learn like humans; they’re building on years of experience actually making it happen.”
Rebecca Kaden, managing partner at Union Square Ventures, added: “Robots become so much more useful when they can instantly pick up a new task and get to work right away.
“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. Whether this is for a robot in the factory or at home, we need general-purpose AI which can adapt quickly in the same way that humans can.”