MIGHTYFP™ - A SWAPC2-OPTIMIZED AI ACCELERATOR FOR DEFENSE AND AEROSPACE APPLICATIONS

The DAF increasingly relies on AI technologies such as machine-based reasoning, predictive analytics, and cognitive human-machine teaming for strategic and tactical advantages. However, current microchips are unfit for AI at scale and unable to handle growing data and model complexity. The DAF needs efficient AI chips for faster training, accurate inference, and lower energy use to maintain decision-making superiority. Arith Inc.’s mightyFP™, a high-performance AI chip accelerator, meets this need, strengthening the U.S. edge in the global AI and semiconductor races.mightyFP™ uses patented fused operations to deliver 6X faster AI training, 60X faster inference, 8X higher throughput, 90% less energy consumption, and 87% memory savings, ideal for real-time missile guidance and targeting applications. In Phase I, Arith validated the demand for AI-driven missile autonomy and computer vision for targeting on SWaPC2-constrained platforms.In Phase II, to support TACPs and broader missile-based autonomous targeting missions, Arith will integrate mightyFP™ into edge computing systems for real-time object detection, classification, and tracking in contested and disconnected environments.mightyFP™’s standard-based interoperability and modularity ensure secure and scalable adoption within existing avionics and AI pipelines, and broad compatibility with any microprocessor architecture and trusted foundry processes. It has proven commercial viability in data centers and AI-powered AR glasses, reducing silicon usage by 80%, targeting the $502 billion AI hardware market by 2033.