• Building on NTU partnership, forged multi-year collaboration with ITRI to advance RRAM to 22-nm
• 22-nm RRAM aims to define new performance-efficiency frontier for neuromorphic and in-memory computing
• Established world-class Technical Advisory Board to guide transition to commercial rollout

dorsaVi announced the commencement of a 22-nm RRAM fabrication and co-design program, undertaken in collaboration with Nanyang Technological University (NTU) and the Industrial Technology Research Institute (ITRI), to advance the company's proprietary RRAM architecture. The 22-nm RRAM program represents a strategic inflection point, enabling a customisable memory and in-memory computing platform that is configurable across a wide range of use cases and operating requirements. This move is driven by increasing demand for ultra-low-power, low-latency and application-tunable computing solutions in robotics, wearables, industrial automation, and medical devices. The collaboration brings together dorsaVi's proprietary sensor and process-in-memory IP, NTU's advanced RRAM device and materials research, and ITRI's expertise in RRAM device and circuit design, enabled by a world-leading commercial foundry service. The 22-nm RRAM platform is being developed to address multiple high-growth end markets where ultra-low latency, energy efficiency, and on-device intelligence are critical, including robotics, wearables, electric vehicles, industrial automation, neuromorphic computing, and ultra-edge AI. Additionally, dorsaVi established a world-class Technical Advisory Board to guide the company's transition from research-grade validation of its next-generation RRAM-enabled neuromorphic sensor and robotics platforms into full-scale commercial rollout.

The 22-nm RRAM program targets lower write voltage, reduced latency, improved reliability, and support for compute-in-memory operation relative to the current 40-nm RRAM devices. These improvements are intended to deliver higher performance per watt and more efficient AI and neuromorphic processing in power- and latency-constrained ultra edge applications, including robotics, drones, and autonomous platforms.

dorsaVi believes its transition to an advanced 22-nm architecture is aligned with addressing the constraints of traditional computing architectures by enabling more in-memory designs that support computing in-memory. The 22-nm program is intended to support improved power efficiency, tighter coupling of memory and logic for faster local processing, and integration of in-memory compute structures that reduce data-transfer overhead.