NVIDIA Selects Micron for SOCAMM Memory Rollout

Jul 23, 2025
NVIDIA plans to deploy 600,000 to 800,000 SOCAMM modules in 2025, with Micron leading the production efforts, according to a press release by DIGITIMES ASIA.

NVIDIA is set to deploy between 600,000 and 800,000 SOCAMM modules in 2025, marking a significant step in the memory market. This initiative, announced in a press release, positions SOCAMM as a potential successor to high-bandwidth memory (HBM), with the company's GB300 "Blackwell" platform among the first to adopt this technology.

Micron Technology has emerged as the first memory maker approved for volume production of SOCAMM, outpacing competitors Samsung Electronics and SK Hynix. SOCAMM is designed for low-power, high-bandwidth AI computing, offering a 2.5x increase in bandwidth and a one-third reduction in size and power consumption compared to traditional RDIMM modules.

While the initial deployment will focus on AI servers and workstations, the inclusion of SOCAMM in the Digits AI PC indicates broader ambitions for the consumer market. This move could reshape dynamics in the substrate sector, as the module requires custom-designed PCBs, creating new demand opportunities for substrate suppliers.

We hope you enjoyed this article

Consider subscribing to one of our newsletters like Silicon Brief or Daily AI Brief.

Also, consider following us on social media:

AI Chips & Datacenters AI Brief AI Brief (X)
Free newsletter

Silicon Brief

Weekly coverage of AI hardware developments including chips, GPUs, cloud platforms, and data center technology.

AI Token Webinar

Whitepaper

Tensordyne Napier: What If One Rack Could Do the Work of Nine?

Tensordyne

This Tensordyne whitepaper presents Napier, an inference-focused AI processor and rack-scale system based on the company’s TDN Math logarithmic number system. It examines infrastructure requirements for large mixture-of-experts and agentic models, compares major inference architecture approaches, and details the TDN AIP processor, TDN72 pod, TDN Link fabric, and Napier Ultra configuration. The paper reports simulation-based performance, cost, and accuracy-validation results, including Tensordyne’s projected comparison of one Napier rack with a nine-rack Nvidia Rubin plus Groq deployment; the chip is reported as taped out and in fabrication.

Read more

You may also like

Free, six days a week

Daily AI Brief: the AI news that matters, in your inbox.