10172871Accelerating Customer Adoption of Photonic AI through FPGA DemonstrationClosedFast Start ResponseInnovate UKLumiAIres is pioneering a new class of Artificial Intelligence (AI) hardware designed to dramatically reduce the energy and latency costs of machine learning. Our technology uses light rather than electricity to perform AI computations, enabling ultra-fast, low-power processing that could transform the way artificial intelligence is deployed across industries. This project will deliver the first live demonstrator of our unique neuromorphic photonic architecture through an object recognition system. Instead of relying on traditional digital processors like GPUs or TPUs, the system will use a novel analogue approach that emulates how neural networks operate in the brain---only with light. The chip architecture is inspired by the physics of wave interactions and is designed to form dynamic, self-organising pathways as it processes information. In this early phase, we will first simulate the architecture as a digital prototype, allowing us to test and validate it through a cloud-based Application Programming Interface (API). This enables rapid iteration and user testing in real-world conditions. Then, we will map the system to a Field Programmable Gate Array (FPGA), a flexible hardware platform that mimics the functionality of the future chip. This will allow us to demonstrate ultra-fast object recognition in real time, on-device, without relying on cloud inference or power-hungry accelerators. The project aims to achieve three outcomes: 1. Build and test a working object recognition demonstrator running on a digital emulation of our chip. 2. Deploy the same architecture on an FPGA to show edge-hardware-level performance and latency. 3. Use this demonstrator to engage early design partners and prepare for the next phase of physical chip fabrication. The impact of this project could be far-reaching. AI workloads are placing increasing strain on global energy infrastructure, especially in data centres and edge devices such as mobile systems and satellites. By offering a radically more efficient and scalable computing substrate, this innovation supports the UK's goals in sustainable digital technologies, green compute, and sovereign hardware capability.