How Will dorsaVi’s 22 nm RRAM Revolutionize Wearable Tech and AI?
dorsaVi has launched an evaluation program to scale its oxide-based RRAM technology from 40 nm to 22 nm, aiming for significant gains in density, speed, and energy efficiency. This move could enhance its wearable sensor devices and embedded AI applications.
- 40 nm RRAM baseline shows strong performance with 50–200 ns switching and >10 million cycles
- 22 nm evaluation targets ~3.3× higher bit density and 40–60% energy reduction per bit
- Projected sub-100 ns switching speed and lower write voltage (~1.5 V) at 22 nm
- Plan to tape-out 22 nm test chip with TSMC pending final validation
- Advances support wearables, clinical biosignal processing, and robotics applications
Scaling Up RRAM Technology
dorsaVi Limited, an ASX-listed leader in wearable sensor technology, has announced the commencement of a formal evaluation to scale its oxide-based resistive random-access memory (RRAM) from the 40 nm process node to an advanced 22 nm node. This initiative builds on the company’s already validated 40 nm platform, which has demonstrated robust switching speeds, endurance, and retention, serving as a strong foundation for next-generation embedded memory solutions.
The move to 22 nm is designed to deliver meaningful improvements in memory density, energy consumption, and switching speed. Early analysis suggests a roughly 3.3-fold increase in bit density, a 40–60% reduction in energy per bit, and switching speeds dropping below 100 nanoseconds. These enhancements are expected to support dorsaVi’s roadmap for more compact, efficient, and responsive wearable devices and AI-driven systems.
Why 22 nm? Balancing Innovation and Manufacturability
The choice of the 22 nm node reflects a pragmatic balance between cutting-edge scaling and manufacturing maturity. Unlike smaller nodes such as 16 or 14 nm, 22 nm offers improved lithography and process uniformity without the yield and complexity challenges that can arise at more advanced nodes. This ensures reliable device performance and integration with mixed-signal peripherals critical for compute-near-memory and reflex architectures.
Moreover, the 22 nm node maintains the voltage headroom necessary for reliable RRAM operation, with projected write voltages around 1.5 V, down from approximately 1.8 V at 40 nm. This reduction contributes to lower power consumption and simplifies input/output design, key factors for wearable and edge AI applications where energy efficiency is paramount.
Applications in Wearables and Robotics
dorsaVi’s technology is already gaining traction in clinical and industrial safety markets, with its ViMove+ system used for real-time biosignal processing such as electromyography (EMG) and electrocardiography (ECG). The transition to 22 nm RRAM is expected to enable longer device wear times, faster data reporting, and smaller form factors, enhancing user experience and clinical workflows.
In robotics, the improved memory technology supports smaller reflex modules with tighter control loops and lower power draw, facilitating better integration into constrained spaces like joints or end-effectors. This advancement could improve the performance and reliability of collaborative robots, exoskeletons, and safety-critical systems.
Next Steps and Market Implications
Subject to successful device- and array-level validation, dorsaVi plans to proceed with a 22 nm test-chip tape-out using TSMC’s 22 nm process. This milestone will mark the transition from evaluation to fabrication, a critical step toward commercial deployment. The company also continues to explore complementary technologies to strengthen its technical stack at this node.
Overall, this development signals dorsaVi’s commitment to innovation in embedded non-volatile memory and edge AI, positioning it well to capitalize on growing demand for advanced wearable sensors and intelligent robotics solutions.
Bottom Line?
dorsaVi’s 22 nm RRAM evaluation sets the stage for faster, denser, and more energy-efficient memory that could redefine wearable and AI-driven device performance.
Questions in the middle?
- When will dorsaVi complete validation and confirm the 22 nm tape-out timeline?
- How will the company manage commercial terms and production scale with TSMC?
- What competitive advantages will dorsaVi’s 22 nm RRAM offer against other emerging memory technologies?
