Global debut! Intel unveils ZAM memory prototype: single chip 512GB, half the power consumption, directly competing with HBM

Intel and SoftBank's jointly developed next-generation AI memory technology, Z-Angle Memory (ZAM), made its global debut, showcasing significant performance advantages as a new technology aiming to challenge the dominance of the high bandwidth memory (HBM) market. The product adopts a vertical stacking architecture, which is expected to greatly increase capacity while reducing power consumption. On Wednesday, according to Wccftech reports, early data shows that ZAM can reduce power consumption by 40% to 50%, with a single chip capacity of up to 512GB, and simplifies the manufacturing process through Z-Angle interconnect technology. These features make it a potential solution for current AI applications facing power consumption bottlenecks and supply chain shortages. SoftBank subsidiary SAIMEMORY first showcased the ZAM prototype on February 3 at the Intel Connection Japan 2026 event. Intel revealed in a blog post that the prototype is planned for launch in 2027, with full commercialization expected by 2030. Intel’s Chief Technology Officer for Government Technology, Intel Fellow Dr. Joshua Fryman, attended the event. SAIMEMORY was jointly established by SoftBank, Intel, and the University of Tokyo in December 2024, officially began operations in June 2025, and is currently led by President and CEO Hideya Yamaguchi. Vertical stacking breaks heat dissipation bottleneck According to PC Watch, traditional memory uses a planar stacking structure, but this design is reaching its limit due to power consumption and heat dissipation constraints. The current design has pushed 16-layer stacking close to the maximum, with 20 layers considered the upper limit. ZAM is named after the Z-axis and employs a vertically stacked chip design. PC Watch points out that compared with conventional DRAM, this design promises lower power consumption, higher capacity, and wider bandwidth. Through vertical stacking, heat generated by each chip can be conducted evenly upwards, solving the long-standing heat dissipation challenges of planar stacking. Wccftech cites Intel as saying that the main advantage of this architecture lies in its superior thermal management capability. Technical path aims to fill gap between HBM and DDR This new memory product is expected to use Intel’s next-generation DRAM bonding (NGDB) technology. According to information released by the Sandia National Laboratory in January, current high bandwidth memory often sacrifices capacity and other performance metrics to achieve greater bandwidth. NGDB technology aims to eliminate such trade-offs, bridging the gap between HBM and traditional DDR DRAM, while significantly improving energy efficiency. According to SoftBank's press release, the ZAM project's prototype product is expected to be completed in the fiscal year ending March 31, 2028, with commercialization targeted for the 2029 fiscal year. Intel Fellow Dr. Joshua Fryman stated that standard memory architectures cannot meet AI demands, and Intel's new architecture and assembly method improves DRAM performance while reducing power consumption and cost. PC Watch reports that SAIMEMORY emphasizes its strong partnerships, including collaborations with SoftBank and Intel as well as a network of domestic and international investors and supply chain partners. This suggests that Intel may not be SAIMEMORY’s sole global partner. The collaboration between SoftBank and Intel aims to carve out a new path in a HBM-dominated market, directly addressing the power consumption bottlenecks and supply chain constraints faced by current AI applications. Risk Warning and Disclaimer The market entails risks, and investment requires caution. This article does not constitute personal investment advice nor does it take into account individual users’ specific investment objectives, financial situation, or needs. Users should consider whether any opinions, viewpoints, or conclusions in this article are suitable for their particular circumstances. Investing based on this is at your own risk.