Rambus Launches Chipset for AI Server Memory Modules – Investing.com Sweden

Rambus Inc. Has announced a new chipset designed to enable low-power, high-performance memory modules for artificial intelligence server platforms, marking a significant step in the company’s strategy to support evolving data center workloads driven by AI.

The company revealed the SOCAMM2 (Small Outline Compression Attached Memory Module) chipset on April 22, 2026, which is engineered to support JEDEC-standard LPDDR5X-based memory modules operating at speeds up to 9.6 Gb/s. This launch represents the first product in a planned family of LPDDR-based server module solutions aimed at improving power efficiency, bandwidth, and form factor for AI infrastructure.

The SOCAMM2 chipset includes newly developed voltage regulators and an SPD (Serial Presence Detect) hub, integrating Rambus’ expertise in signal and power integrity to enhance performance, reliability, and interoperability in advanced memory architectures. These components are critical for managing power delivery and telemetry in compact, high-density server environments.

According to Rambus, the SOCAMM2-based memory modules are designed to address growing demands in AI data centers by delivering high bandwidth with lower power consumption in a modular, serviceable form factor that optimizes board space. The company positions this technology as an innovative approach to scaling memory performance while meeting strict efficiency requirements in modern AI platforms.

This announcement extends Rambus’ existing portfolio of memory interface chipsets for JEDEC-standard DDR5 and LPDDR5 modules, reinforcing its role as a provider of comprehensive silicon IP for data-intensive applications. The company stated that the SOCAMM2 launch reflects ongoing collaboration with industry partners to develop memory solutions tailored to the specific needs of AI workloads, including large-scale model training and inference.

Rambus emphasized that the SOCAMM2 chipset is the initial step in a broader roadmap focused on LPDDR-based server modules, signaling continued investment in memory technologies that balance performance with power efficiency for next-generation AI systems.