Luffy AI secures £8.1 million to integrate self-tuning AI into electric motors globally.

      Luffy AI, an Abingdon-based startup developing what it refers to as neuroplastic AI for real-time management of physical machines, announced on Tuesday that it has secured £8.1 million in a Series A funding round.

      The round was led by BGF, claiming to be the most active equity investor in the UK and Ireland, with the funds designated for transitioning pilot projects into commercial applications. Luffy is a true spinout; its founders, Dr. Matthew Carr and Dr. Alex Meakins, are former nuclear physicists from the UK Atomic Energy Authority, and the company operates on the Culham Campus near Oxford, a location primarily recognized for fusion research rather than the university deeptech spinouts that are increasingly emerging from there.

      The company's approach addresses a challenge that the broader AI surge has largely overlooked. Large language and image models depend on extensive data, cloud computing, and ongoing connectivity, which are not appropriate for machines like pumps or conveyor belts on a factory floor. This stands in stark contrast to the physical-AI investments that have attracted significant funding in the UK, such as Wayve’s $1 billion round.

      Luffy proposes a novel network design. Its sparse neural networks are trained in simulations to bypass the necessity for large training datasets that traditional deep learning requires. They are then fine-tuned using real machines, an approach that the company claims can be up to 400 times more efficient than conventional deep learning.

      The architecture is designed to be compact enough to run directly on the hardware. Since the models refine themselves using real-time feedback instead of requiring cloud retraining, Luffy asserts that they can operate at the edge, continuously adapting to whatever they are managing.

      The initial focus is on electric motors. Electric motors account for about half of global electricity consumption, many of which operate inefficiently. Luffy is implementing its models in motor control and variable-frequency-drive applications for industrial pumps, fans, and conveyors.

      The commercial potential promotes a plug-and-play solution. According to the company, adaptive control enables a motor to adjust according to its load and operational conditions, reducing energy consumption, expediting commissioning time, and enhancing performance without needing a specialized engineer on-site.

      “AI has revolutionized language and image generation but has yet to significantly impact industry beyond predictive maintenance and dashboards,” stated Carr, Luffy’s co-founder and CEO. He emphasized that factories and motors require AI that is "small, fast, and adaptive in real time," rather than reliant on cloud technology and excessive data and computational resources.

      Joining BGF in this investment round were MIG Capital, a Munich-based deep-tech investor through its MIG Fonds, as well as existing supporters Bow Capital, Chrysalix, Momenta, and UKI2S. Kate Ronayne, an early-stage investor at BGF, noted that Luffy is “disrupting an industry norm that has persisted for a century” by integrating specialized AI directly into physical systems and minimizing reliance on expert engineers during commissioning.

      For MIG, efficiency was as attractive as ambition. “Luffy operates with significantly less data and compute, which is exactly what makes AI feasible inside physical machines,” remarked Dr. Nicolas Rose-André, an investment manager at the firm, highlighting the massive electricity consumption of motors as an opportunity in its own right.

      Luffy did not disclose its valuation for the funding round or its revenue to date. The financing will advance its proofs of concept and pilot projects toward partnerships with larger industrial brands, and the company envisions the same control technology being applied to robotics and drones, thermal process control, and other physical-AI applications.

      This ambitious goal, pursued on the basis of a modest funding round, hinges on whether a self-tuning motor can demonstrate its efficacy on actual factory floors rather than solely in simulations.