- 01DorsaVi starts targeted RRAM energy tests.
- 02Aimed to set baseline energy for low-power hardware; informs CIM.
- 03Supports ultra-edge AI via local memory near sensing.
dorsaVi (ASX: DVL) has begun targeted electrical testing of its engineered resistive random-access memory (RRAM) structures to measure the energy required to update memory states.
The work is intended to establish an initial energy-performance benchmark for dorsaVi’s low-power hardware pathway before more comprehensive testing with its RRAM Validation Chip.
The results will help inform further optimisation, reliability work, array-level evaluation, and future Compute-in-Memory (CIM) testing as dorsaVi develops more integrated hardware capability.
The company sees physical AI, including advanced robotics, autonomous systems and adaptive wearables, as a potential application area where local memory and processing could reduce the need to send all raw sensor data to a distant processor.
Device Energy Benchmark
The evaluation focuses on device-level energy associated with changing RRAM memory states, a characteristic that may become increasingly relevant where memory is updated frequently close to sensing functions.
dorsaVi cautions that the measurements will not represent the full power consumption of an AI system, which would also include sensing, logic, communications, actuation, and other components.
Any eventual system-level benefit from lower-energy RRAM updates will depend on factors including array design, integration approach, compute architecture and workload, with further validation still required.
The planned Validation Chip program will extend the work towards array-level operation and future CIM evaluation alongside further reliability and workload-level assessment.
Robotic Hands Illustrate Opportunity
“Advanced robotic hands are a powerful reference application because they show why memory, processing, and energy efficiency need to move close to the point of contact,” group chief executive officer Mathew Regan said.
“Our study is focused on the energy efficiency of our RRAM structures, and how that may support the next generation of ultra-edge hardware.”
Unlike conventional industrial grippers designed for repeatable movement in structured environments, next-generation hands may need to manipulate irregular parcels, flexible materials, delicate components, tools, and everyday objects.
Tactile sensing across fingers and the palm could generate pressure, contact, force, and slip data.
Additionally, local or near-sensor memory could retain selected context close to where it is created before surrounding logic supports interpretation and response.
Broader Ultra-Edge Applications
dorsaVi has identified other illustrative AI applications including autonomous vehicles and mobile machines, industrial automation, wearable sensing, warehouse and logistics robots, and medical or surgical robotics.
Although their requirements differ, the common premise is that lower-latency and more energy-efficient intelligence may be valuable where power, space, and thermal capacity are constrained at the ultra-edge.
The RRAM technology would provide the memory function within a broader architecture, with local processing and responsive action requiring surrounding logic and potentially later CIM or neuromorphic functions.
dorsaVi’s current testing is therefore aimed at establishing one foundational energy metric before the company moves into more integrated validation of its sensing, memory, and processing pathway.
Get the wire before the market opens.
The ASX small-cap stories that matter, filed before 9am AEST. Curated by the Small Caps desk.