Marvell Technology, Inc. announced on June 17, 2025, the launch of the industry’s first 2nm custom Static Random Access Memory (SRAM), designed to enhance the performance of custom XPUs and devices powering AI and cloud data centers. This breakthrough aims to optimize memory hierarchy for next-generation AI infrastructure.
Marvell introduces the first 2nm custom SRAM for AI and cloud infrastructure.
Delivers 6 gigabits of high-speed memory at up to 3.75 GHz.
Reduces power consumption by up to 66% compared to standard SRAM.
Saves up to 15% of die area, enabling more compute cores or smaller chips.
Part of Marvell’s custom silicon platform, including CXL and HBM technologies.
Supports hyperscalers in scaling AI infrastructure efficiently.
Marvell’s 2nm custom SRAM delivers up to 6 gigabits of high-speed memory, achieving the industry’s highest bandwidth per square millimeter. "Custom is the future of AI infrastructure. The methodologies and technologies used by hyperscalers today to develop cutting-edge custom XPUs will percolate to more customers, more classes of devices, and more applications," said Will Chu, senior vice president of Custom Cloud Solutions at Marvell. This innovation allows chip designers to reclaim up to 15% of die area, enabling more compute cores, expanded memory, or reduced chip size and cost.
Operating at speeds up to 3.75 GHz, the SRAM reduces standby power consumption by up to 66% compared to standard on-chip SRAM at equivalent density. "Memory remains one of the biggest challenges for AI clusters and clouds. These systems need as much memory as they can get as fast as they can," said Alan Weckel, co-founder of the 650 Group. Marvell’s innovative data path and architecture ensure high performance while addressing the power demands of AI workloads.
The 2nm SRAM is a key component of Marvell’s custom technology platform, which includes CXL for memory expansion and custom HBM for increased memory capacity. Built on TSMC’s 2nm process, the platform integrates advanced IP like PCIe Gen 7, die-to-die interconnects, and silicon photonics to support custom AI accelerators and networking solutions. This holistic approach addresses the post-Moore’s Law era, where transistor scaling is less viable, and custom silicon drives performance and efficiency.
Marvell’s 2nm custom SRAM sets a new standard for AI and cloud infrastructure, offering unmatched efficiency and flexibility. By enabling hyperscalers and developers to optimize chip designs, Marvell strengthens its position as a leader in custom silicon for next-generation data centers.
To deliver the data infrastructure technology that connects the world, we're building solutions on the most powerful foundation: our partnerships with our customers. Trusted by the world's leading technology companies for over 25 years, we move, store, process and secure the world's data with semiconductor solutions designed for our customers' current needs and future ambitions. Through a process of deep collaboration and transparency, we're ultimately changing the way tomorrow's enterprise, cloud, automotive, and carrier architectures transform—for the better.
TechIntelPro empowers B2B technology decision-makers with the latest industry trends, deep insights, leadership perspectives and actionable intelligence to drive both organizational and professional milestones—and thrive on the tech pulse.