Intel Corp. announced a collaboration with SAIMEMORY, a subsidiary of SoftBank Corp., to develop a new stacked memory technology known as Z-Angle Memory (ZAM). The initiative targets next-generation memory architectures to support the rapidly rising demands of AI and high-performance computing workloads.
SAIMEMORY, based in Tokyo, is developing a stacked DRAM architecture designed to exceed current high-bandwidth memory capabilities by increasing memory capacity, reducing power consumption, and advancing packaging techniques. The approach aims to address memory bottlenecks that constrain scaling in large AI systems. Under the agreement, Intel will act as a technology, innovation, and standards collaborator, while SAIMEMORY will lead technology development and commercialization. Operations are scheduled to begin in Q1 2026, with prototypes expected in 2027 and commercialization targeted for 2030.
The ZAM effort builds on prior research funded by the U.S. Advanced Memory Technology (AMT) R&D Program, managed by the Department of Energy and the National Nuclear Security Administration through Sandia National Laboratories, Lawrence Livermore National Laboratory, and Los Alamos National Laboratory. Intel contributed early proof points through AMT-funded work and its Next Generation DRAM Bonding (NGDB) initiative, which demonstrated higher DRAM density and bandwidth with lower latency and energy consumption. Intel will now apply these learnings to support the ZAM program alongside SAIMEMORY.
- ZAM focuses on stacked DRAM architectures beyond current HBM designs
- Intel provides standards, validation, and technology collaboration; SAIMEMORY leads commercialization
- Program timeline: operations in Q1 2026, prototypes in 2027, commercialization by 2030
- Builds on AMT and NGDB research validating higher density, bandwidth, and energy efficiency
- Strengthens U.S.–Japan collaboration in advanced memory technologies
“Intel’s Next Generation DRAM Bonding initiative has demonstrated a novel memory architecture and revolutionary assembly methodology that significantly increases DRAM performance, reduces power consumption, and optimizes memory costs,” said Dr. Joshua Fryman, Intel Fellow and CTO of Intel Government Technologies. “Standard memory architectures aren’t meeting AI needs, so NGDB defined a whole new approach to accelerate us through the next decade.”
🌐 Analysis
The ZAM collaboration reflects growing industry focus on memory-centric innovation as AI accelerators face bandwidth, capacity, and power limits with existing HBM roadmaps. Intel’s role positions it alongside other ecosystem players exploring advanced stacking and bonding approaches, while SoftBank-backed SAIMEMORY adds a commercialization path that bridges national-lab research and large-scale manufacturing. The program also aligns with broader efforts in the U.S. and Japan to secure advanced semiconductor supply chains for AI and HPC systems.