• Validated reflex loops completed in <1 μs, exceeding biological-grade requirements
• Each reflex confirmed at ~1 nJ per inference, with standby leakage negligible in the nW-μW range
• Demonstrated orders-of-magnitude faster responses and dramatically lower energy consumption than leading digital edge AI systems

dorsaVi Limited (ASX: DVL) has announced a major technical breakthrough in its RRAM Reflex Platform, demonstrating that it can achieve biological-grade reflexes with sub-microsecond latency and nanojoule-level energy use. By benchmarking against state-of-the-art digital edge AI systems, dorsaVi has shown that its RRAM-based platform delivers orders-of-magnitude improvements in both speed and efficiency. The key highlights include: validated reflex loops completed in less than 1 microsecond, each reflex confirmed at around 1 nanojoule per inference, and standby leakage in the nanoWatt to microWatt range. This breakthrough positions RRAM as the nervous system for next-generation robotics and biomedical systems, collapsing the entire sense-compute-act chain into a single fabric and removing the bottlenecks inherent in digital systems. While digital edge AI will remain critical for high-level reasoning, the RRAM Reflex Platform sets a new benchmark for how machines can safely and intelligently interact with the physical world, enabling practical applications such as robotic grippers that release fragile objects before damage occurs and prosthetic limbs that can recoil from excessive heat or force to protect wearers.

With engineering validation complete and benchmark superiority confirmed, the next updates will focus on robotic evaluation and commercial translation.