• Machenta raised $40 million in seed funding led by Creandum and Union Square Ventures.
• It says robots can learn tasks from one demo using a model trained entirely in simulation.
• The startup remains focused on research as it takes on multibillion-dollar rivals.

A London-based robotics startup is betting it can teach robots new tricks without collecting the mountains of real-world data its much bigger rivals are chasing.

Machenta, founded by researchers from Imperial College London, has raised a $40 million seed round to build an AI model that can teach robots new physical tasks after seeing them performed just once, the company exclusively told Business Insider. Venture capital firms Creandum and Union Square Ventures co-led the round. Machenta declined to disclose its valuation.

While rivals building general-purpose AI models for robots are scrambling to amass real-world data, Machenta is taking a contrarian approach. Edward Johns, the startup's CEO and cofounder, and an Imperial professor, said the company trains its model in computer-generated environments that mimic how objects move and interact, without using a "single pixel of real-world data."

Once trained, the model can direct a robot to perform a new task, such as tidying a table or putting objects into a box, after watching it done just once, without further training, Johns said. In a video shared with Business Insider, three robots pack and close a box on an assembly line, a task Johns said they learned this way.

The startup has 11 employees and remains focused on research, with no near-term plans to commercialize its technology. Its model can learn some tasks immediately, but more precise tasks and complex movements remain challenging.

Machenta faces rivals with multibillion-dollar valuations in one of robotics' most competitive and well-funded fields. Its California-based competitors include Physical Intelligence, valued at roughly $11 billion, and FieldAI, which is raising funding at a $10 billion valuation, Business Insider reported this month.

Another rival, Skild AI, which is valued at over $14 billion, is already selling technology with similar capabilities. In August, Skild unveiled a model that it said could perform previously unseen tasks lasting up to 10 minutes after a single video demonstration. Last month, it said that it had reached $100 million in annual recurring revenue from more than 60 customers.

From research lab to startup

Six years ago, Johns began investigating whether robots could learn new physical tasks as quickly as humans at his Imperial College lab.

In the lab's early experiments, researchers physically guided a robot through a task, such as picking up a spatula and scooping an object. The robot then spent about 30 minutes collecting additional data about the object and its surroundings before it could perform the task on its own.

Johns and former Ph.D. student Vitalis Vosylius spent the next few years working to eliminate that extra step, founding Machenta late last year after seeing early signs that their model could perform new tasks immediately after a demonstration.

Companies are increasingly buying and collecting data to teach robots how to interact with the real world. But Johns believes simply collecting more data isn't the answer. It also remains unclear how much training data will be needed to make general-purpose robots a reality. Johns said that Machenta is taking on the harder challenge of designing algorithms that can generate synthetic data useful for teaching robots physical tasks.

Machenta plans to use the funding to expand its team and train its model with more computing power and simulated data, hoping to teach robots to perform more precise and complex tasks.

Have a tip? Contact Rya Jetha via email at rjetha@businessinsider.com or Signal at rjetha.07. Use a personal email address and a non-work device; here's our guide to sharing information securely.

Ukraine’s six-wheeled Frodo ground robot fired on targets during a demonstration in Kharkiv Oblast for the Pomsta (Revenge) Border Guard Brigade, a border guard detachment posted on Facebook. Frodo made its public debut only on 30 September.

Ukraine had contracted more than 22,000 ground robots for 2026 by July, mostly logistics carriers, with combat models the smallest group, the Defense Ministry said. Frodo’s first reported mission was an evacuation. The demo shows its armed version to a fighting brigade, the kind of robot Ukraine has ordered least.

What the armed version carries

The platform completed its tasks at a training ground in near-combat conditions, said the 3rd Border Guard Detachment named after Hero of Ukraine Colonel Yevhenii Pikus. The Buria turret mounted on Frodo can fire either a Mk 19 automatic grenade launcher or a Browning M2 heavy machine gun. It engaged targets at up to 1.3 km (0.8 mile), the detachment said.

The 47-kg (104-lb) module has a ballistic calculator, stores up to 256 targets, aims in three seconds, and has a stated autonomous run time of up to 48 hours, the post said. Frodo is rated for up to 150 km (93 miles) on paved roads and 75–80 km (47–50 miles) over rough terrain, the post added. A Frontline Robotics Zoom reconnaissance drone flew reconnaissance during the test.

The developer, Frontline Robotics, describes its core philosophy as arriving fast, doing the job, and leaving very fast—hit and run, the detachment’s post said. Familiarizing the brigade’s soldiers with such systems aims first and foremost to preserve personnel’s lives and health, the detachment said.

A robot first reported as a medevac

Frontline Robotics told Militarnyi at the unveiling that Frodo had passed Defense Ministry codification and was built for logistics, evacuation, and strike missions. The robot reaches 40 km/h (25 mph) and carries up to 200 kg (441 lb) of payload, or 400 kg (882 lb) at most, the company said. A single soldier can swap its batteries.

The 92nd Assault Brigade used an early test unit to evacuate a wounded soldier, Frontline Robotics told Militarnyi. An operator, callsign Tango, named the robot’s speed and its ability to handle complex maneuvers as its main advantages, Militarnyi reported.

It’s getting better all the time (Better, better, better) It’s getting better all the time I admit it’s getting better –The Beatles In an X post, Marko Jukic, senior analyst at Bismarck Analysis, lamented: “Enshittification” is sort of a “safe horny” term for what is really just outright impoverishment and declining living standards due to […]

The post US enshittification vs Chinese involution appeared first on Asia Times.

What once felt like science fiction, an AI robot Olympics, now looks more possible as robotics improves. Reports mention sprinting, soccer, table tennis, gymnastics, and other tests. The World Humanoid Robot Games in Beijing brings in more than 2,000 robots for 51 events. That scale shows how fast robot sports are growing.

Still, could robots run something like their own Olympics one day? It could happen. But getting a world wide version would mean big upgrades in tech, clearer event rules, stronger safety, and a solid way to organize it across countries.

Robots are already competing in sports

The idea is now more than just theory. The World Humanoid Robot Games has made a global stage where robots take part in sports and show what they can do in motion. In August 2026, a Chinese humanoid robot named Tiangong Ultra is said to have run the 100-meter dash in 8.86 seconds. That time beats Usain Bolt's well known mark of 9.58 seconds.

These events let teams study balance and how a machine moves. For viewers, it is a fun way to watch progress in robotics unfold.

What would an AI robot Olympics look like?

A future AI robot Olympics might look like the human Games, but add contests made for machines.

In the main lineup, there could be robotic sprint races, robot jumping events, robot swimming, robot soccer, robotic weightlifting, and robot gymnastics. Alongside those, there could also be robot-only events that check battery use, fine control, path finding through obstacles, long mechanical runs, gripping and moving objects, and doing tasks with less outside help.

Robots would not be judged in the same way as people. Some scores could come from the kit itself, like the sensors, the motors, and the software setup. Organizers might also ask how well the system handles shifts in its surroundings, or how steadily a robotic hand finishes tricky work.

With this setup, a robot Olympics would feel more like a tech contest than a simple sports meet. It could draw teams from around the world to prove who builds the better machine.

AI would be just as important as hardware

A big contrast between sports played by people and robot contests is how much artificial intelligence is involved. A robot’s body still matters, but it is also tested on how it reads the world, avoids trouble, and chooses what to do next.

The International Olympic Committee has pointed out that AI matters more and more. It did this with the Olympic AI Agenda, and the plan talks about using AI in a way that stays focused on people throughout the Olympic Games.

So a coming robot event could link AI work with real athletic effort. The contest could set up tasks where deciding on the fly is just as critical as raw speed.

There are still major challenges

Robots have made real strides, but they still cannot fully take over normal sports run by people. A few humanoid machines can sprint at high speed. Yet they often trip on simple tasks like slowing down, staying upright, or getting back up after a fall. In recent events, there were crashes that showed gaps in current engineering.

If robots were to compete regularly, clear rules would be needed. People would have to decide what kind of machines are allowed. Would it be remote control only, full autonomy, or some kind of mix with assistance? There would also need to be set limits for the gear and the code. Battery limits and safety rules would matter too. The goal would be to keep the games fair even in PH Bingo on GameZone.

Could it become a major global event?

It feels more likely now than it did a few years back. More countries are joining in already, and many robotics events are getting tougher and more advanced. If the tools keep getting better, an event built around AI robots could bring in makers, schools, major tech firms, and robotics teams from many places.

For people who like tech and digital games, this kind of change also shows how fast interactive stuff is moving. If you have been watching new trends in gaming and online entertainment, you might run into sites such as PH Bingo on GameZone. That is one more sign that digital tools keep reshaping how people relax.

An AI robot Olympics plan would still take time, so it is not ready yet. Still, the base work is happening. Today’s robot contests already show that machines can handle more complex physical tasks. As AI models, sensors, power systems, and robot parts improve, the gap between a science demo and a real sport may shrink in a way that draws more attention.

Right now, the traditional Olympics belongs to people. But the growth of robot contests hints that a separate global event for machine athletes may be closer than many expected.