DVL RRAM Development Progress Towards 22nm Technology
| Stock | Dorsavi Ltd (DVL.ASX) |
|---|---|
| Release Time | 5 Dec 2025, 8:39 a.m. |
| Price Sensitive | Yes |
DVL Advances RRAM Tech for Edge-AI Platforms
- RRAM scaling tuned for dorsaVi's neuromorphic portfolio
- Optimized RRAM resistance for 22nm integration
- High switching contrast validates robust data retention
DorsaVi Limited (ASX: DVL) has provided an update on the continued progress of its RRAM (Resistive RAM) development program with Artemis Labs. This work forms a critical foundation for the next generation of dorsaVi's biosensing and edge-intelligent systems, where sensing, memory, and computation are tightly integrated in compact, low-power silicon. The RRAM scaling program is being modified to serve as a compatible memory fabric for the company's newly acquired neuromorphic Processing-in-Memory and Adaptive Interface IP, including the Reflex Engine and intelligent interface layers. Artemis Labs is refining the RRAM material stack and resistance characteristics required for reliable operation at the 22nm node, focusing on engineering an optimal resistance window, maintaining clear separation between low and high resistance states, validating behavior under 22nm-class electrical conditions, and tuning materials and processes to unlock low voltage, low power operation with robust data retention. Initial testing of new proprietary RRAM materials has delivered a strong On/Off switching ratio, providing a clear memory window suitable for advanced-node scaling. This high switching contrast is particularly important for in-memory computing architectures, where memory cells are used not just to store data but also to perform computation. By combining advanced RRAM devices with neuromorphic architectures, dorsaVi is laying the groundwork for future hardware modules where sensors can interpret, learn, and adapt at the edge, supporting energy-efficient edge intelligence in wearables, robotics, and industrial sensing systems.
By tuning RRAM for 22nm integration, the company is positioning its memory technology on the same node class used for modern edge-AI processors, paving the path toward manufacturable neuromorphic silicon that can be deployed in real-world devices.