Machenta Ltd is a London-based frontier AI company building a robot foundation model for Instant Robot Learning, in which a robot shown a task once can perform it immediately without further training. The company was founded in 2025 by Edward Johns, Professor of Robot Learning at Imperial College London, and Vitalis Vosylius, his former PhD student. It builds on roughly six years of one-shot robot learning research at Imperial's Robot Learning Lab, including the Instant Policy paper (ICLR 2025) and the Learning a Thousand Tasks in a Day work published in Science Robotics.

The model is pretrained entirely in simulation using computer-generated environments, without real-world training data. It is reported to learn tasks such as packing, assembly and tidying from a single demonstration, and to run on a range of robots including humanoids. Machenta has no near-term commercialisation plans and no independent evaluations of the model are reported.

In October 2026 the company emerged from stealth with a $40M seed round co-led by Creandum and Union Square Ventures, which Imperial College London described as the largest seed round raised by a UK robotics startup. The company had 11 employees at the time. The funding is earmarked for team growth, computing capacity and simulated training data, and a new office and lab in King's Cross, London. Its valuation has not been disclosed.