aiming to support matrix operations central to AI workloads with increased energy efficiency and performance relative to conventional graphics processing units. Neurophos’ offerings include prototype OPUs and development systems designed to integrate into data center environments, and its platform targets engineering teams and compute-intensive applications that demand scalable AI inference capability. The business collaborates with industry partners and investors to advance its optical compute architecture and develop supporting software stacks and hardware modules for deployment. Neurophos also engages in pilot programs and partnerships to test its photonic processors in real-world compute environments and explore integration pathways with existing infrastructure. Its work leverages research in metamaterials and compute-in-memory design to reduce power consumption and enhance throughput for AI inference tasks.