Intel has announced a strategic collaboration with Saimemory, a SoftBank subsidiary, to advance the development of a stacked DRAM-based memory technology program known as Z-Angle Memory (ZAM). The partnership is aimed at addressing the growing performance and efficiency demands of artificial intelligence (AI) and high-performance computing (HPC) workloads. 

The companies stated that the ZAM program will focus on building next-generation memory architectures capable of supporting the accelerating requirements of AI-driven infrastructure and large-scale data center deployments. 

Partnership Roles and Commercialization Roadmap 

Under the agreement, Intel will serve as a key partner in technology development, innovation, and industry standards, while Tokyo-based Saimemory will lead the commercialization strategy for the ZAM platform. 

The collaboration reflects a longer-term effort to bring alternative stacked DRAM solutions to market as the industry searches for scalable options beyond existing high-bandwidth memory (HBM) architectures. 

Supported by US Advanced Memory R&D Programs 

The ZAM initiative will be supported through the Advanced Memory Technology (AMT) R&D Program, managed by the US Department of Energy and the National Nuclear Security Administration, in coordination with leading national laboratories including: 

• Sandia National Laboratory 

• Lawrence Livermore National Laboratory 

• Los Alamos National Laboratory 

This backing underscores the strategic importance of next-generation memory technologies for both commercial competitiveness and national innovation priorities. 

Intel’s NGDB Architecture as a Foundation 

Intel highlighted that its Next Generation DRAM Bonding (NGDB) initiative has already demonstrated a novel assembly methodology designed to significantly improve memory performance while reducing power consumption and optimizing cost structures. 

Dr. Joshua Fryman, Intel Fellow and CTO of Intel Government Technologies, noted that traditional memory architectures are increasingly insufficient for emerging AI requirements, prompting the need for fundamentally new approaches over the next decade. 

AI Data Center Growth Driving Memory Constraints 

The announcement comes amid intensifying pressure across global memory supply chains, driven by rapid expansion in AI data center construction and infrastructure investment. 

According to TrendForce data cited in industry reports, up to 70% of worldwide memory production in 2026 is expected to be consumed by data centers. Major chipmakers such as Samsung and SK Hynix have also indicated that memory supply constraints could persist through 2027, potentially impacting consumer electronics markets as capacity becomes increasingly allocated toward AI-focused deployments. 

You can review our most recent, relevant findings here:

 https://semiconductorinsight.com/report/high-bandwidth-memory-chips-market/

The Intel-Saimemory partnership signals a significant move toward vertically aligned innovation in advanced memory systems, with stacked DRAM technologies expected to play a critical role in enabling future AI, HPC, and large-scale computing platforms. 

As demand continues to surge for more efficient, higher-performing memory solutions, ZAM represents a long-term effort to reshape the competitive landscape of semiconductor memory architecture through innovation, commercialization readiness, and government-supported R&D alignment.