SoftBank Corp, through its subsidiary Saimemory, and Intel Corporation announced a strategic partnership to develop Z-Angle operation memory technology for AI data centers. The technology provides higher capacity, greater bandwidth, and significantly lower power consumption than conventional memory solutions to address surging demand for high-speed, energy-efficient memory.
The collaboration targets memory needs in AI data centers, where generative AI models require handling massive data throughput while controlling energy costs. High-performance memory represents a critical bottleneck in AI computing, as larger and more complex models demand advanced systems. SoftBank states the new technology addresses these requirements directly through its design specifications.
SoftBank Corp operates as the primary publicly listed unit of SoftBank Group Corp. This announcement extends the relationship between SoftBank Group and Intel, following the group’s mid-2025 agreement to invest $2 billion in the U.S. chipmaker. That investment reflects SoftBank’s confidence in Intel’s position within the global semiconductor and AI ecosystem.
For Intel, the partnership occurs amid efforts to bolster its chip portfolio and build momentum in the competitive AI hardware market. Competitors have pursued AI-related demand aggressively, particularly in data center accelerators and memory solutions. Intel contributes its next-generation DRAM bonding initiative to the effort, enabling advances in memory architecture.
Under the agreement, Saimemory utilizes Intel’s advanced DRAM bonding technology for developing and testing prototypes of the Z-Angle operation memory chips. The companies set early 2028 as the target for initial prototypes. This process aims to demonstrate the technology’s capabilities in capacity, bandwidth, and power efficiency tailored to AI workloads in data centers.
