Nvidia Adds Custom NVHBM to Its NVLink Fusion Roadmap

DigiTimes reports that Nvidia has announced NVHBM, described by the source as a next-generation custom HBM technology. The initiative is presented as part of Nvidia's NVLink Fusion ecosystem for AI infrastructure. The report links the announcement to the growing importance of system-level integration across compute, memory, interconnects, and thermal management as large AI models increase infrastructure requirements. Custom memory architecture could give Nvidia another lever in the design of tightly integrated AI systems. The supplied material does not identify memory suppliers, specifications, production timing, volumes, or commercial products using NVHBM. It is therefore an important demand-and-architecture signal for the HBM ecosystem, not evidence of an immediate change in HBM supply or pricing.