Industry has not yet benefited from AI in the same way as digital sectors. Physical AI models require large amounts of data, embedded applications are constrained by the compute available on-device, and physical assets often lack reliable cloud connectivity. Luffy AI addresses these limitations with a neuroplastic AI stack for real-time adaptive control. The start-up’s efficient neural networks, so-called Sparse Neural Networks, are first trained in simulations and then refined in real-world deployment. Large training datasets are not required. The architecture is lean, energy-efficient and self-refining, avoiding the need for constant retraining from the cloud. According to the company, the technology can be up to 400 times more efficient than traditional deep learning.

Dr Matthew Carr, co-founder and CEO of Luffy AI, said:

“AI has been transformative for language and image generation, but has yet to make a comparable impact in industry beyond predictive maintenance and dashboards. Factories, motors and physical systems need AI that is small, fast and adaptive in real time, not dependent on the cloud and not reliant on enormous amounts of data or compute. At Luffy AI, we have already shown what is possible with AI-based motor control and will use this new funding to expand our delivery and rollout.”

 

Dr. Nicolas Rose-André, Principal at MIG Capital, comments:

“Luffy AI shows that adaptive AI control is possible with significantly less data and compute. That is exactly what makes AI practical inside physical machines. With electric motors consuming around half the world’s electricity, the efficiency potential alone is enormous. We are backing a rare combination of clearly differentiated technology and an outstanding team to bring it to market.”