dorsaVi Characterises RRAM Energy Efficiency for Physical AI

Key points

• RRAM energy testing underway: Dorsavi has commenced targeted electrical measurements to quantify the energy required to update memory states in its engineered RRAM structures.
• Positioned for high-value Physical AI applications: Advanced robotics, autonomous vehicles, and industrial automation are illustrative, high-value markets where local or near-sensor memory and processing may become increasingly important.
• A building block for intelligent Ultra-Edge systems: RRAM provides the memory function within a broader architecture, while associated CMOS logic and, in future, CIM or neuromorphic functions can support local interpretation and system response.

Full summary

Dorsavi Limited (ASX: DVL) has initiated a targeted energy-efficiency evaluation of its engineered Resistive Random-Access Memory (RRAM) structures. This assessment aims to quantify the energy required to update RRAM memory states, providing crucial data for the development of Dorsavi's low-power RRAM-CMOS hardware pathway. The findings will inform further optimization and subsequent evaluations, including reliability, array-level, Compute-in-Memory (CIM), and CMOS integration. The results are intended to support the Company's broader pathway towards more integrated hardware capability. Physical AI is a synergistic use case for Dorsavi's developing sensing, memory, and processing capabilities.