Analysis
Machenta, an 11-person London robotics startup, raised a $40 million seed round co-led by Creandum and Union Square Ventures, Tech Funding News reported on October 27. The company is building a foundation model designed to let robots learn a new task from just one or two real-world demonstrations, without further training.
## What Machenta does Founded in 2025 by CEO Edward Johns, Machenta trains its model entirely in simulation, with the goal of teaching the model "how to learn" rather than memorizing specific tasks. The company says the same model has already learned everyday tasks -- packaging, assembly and tidying -- across several different robot types, including humanoids. Machenta describes itself as a research company with no near-term plans to commercialize, and it has not disclosed hiring plans beyond its current 11-person headcount as it moves into a new office and lab in London's King's Cross. The pitch to investors rests on a narrow but ambitious claim: that a model trained purely on simulated demonstrations can generalize to real hardware it has never touched before, skipping the slow, expensive process of collecting real-world robot data that has historically bottlenecked the category.
## Machenta's funding history This $40 million seed is Machenta's first disclosed institutional round, a notably large check for a company barely a year old with a team in the low double digits. The size reflects how aggressively capital has moved into robot-foundation-model startups over the past year, where seed and Series A rounds have increasingly resembled what used to be Series B territory.
“The company says the same model has already learned everyday tasks -- packaging, assembly and tidying -- across several different robot types, including humanoids.”
## Who invested in Machenta Creandum and Union Square Ventures co-led the round. Creandum partner Kathi Wilhelm said "very few teams in robotics can point to a body of work like this," while USV managing partner Rebecca Kaden framed the bet as backing "general-purpose AI which can adapt quickly in the same way that humans can." Neither firm disclosed additional participants.
## Robotics foundation-model market and competitors Machenta names Physical Intelligence and FieldAI as its principal competitors, and points to Skild AI as a company with similar capabilities -- valued at more than $14 billion after a $1.4 billion Series C led by SoftBank in January 2026.
Physical Intelligence, for its part, is reportedly back in funding talks at a valuation nearing $20 billion. The gap between Machenta's $40 million seed and these competitors' double-digit-billion valuations illustrates how early-stage the company still is relative to a category that has already produced several multi-billion-dollar names in under two years. That gap is also the opportunity Creandum and USV are underwriting: if simulation-only training proves out, Machenta could compress years of data-collection cost that its better-funded rivals have already spent.
## The sim-to-real test Machenta says precise tasks and complex motions remain very difficult for its model, and the company has no timeline for commercialization. The seed capital is earmarked for team growth, more computing power and more simulated training data -- all inputs, not outputs. Whether one-shot learning in simulation actually transfers cleanly to real robot hardware, the "sim-to-real gap" that has stalled prior robotics research bets, is the open question that determines whether Machenta's next round looks like a modest step-up or another Physical-Intelligence-style re-rate.