Primemas Shows CXL Memory Products for Abaco AI System

Aug 24, 2026
Primemas showcased CXL memory products at FMS 2026 and detailed its work with Micron on the Abaco Project, a rack scale compute system planned for Pacific Northwest National Laboratory.

Primemas showcased CXL memory products at Future of Memory and Storage 2026 and detailed its collaboration with Micron Technology on the Abaco Project, announced in a press release. The Abaco Project is a rack scale compute system planned for Pacific Northwest National Laboratory, with more than 100TB of shared and pooled CXL attached memory for scientific AI, HPC, and data workloads.

Primemas said its PMA14 and PMA16 CXL add in cards use PCIe and include 14 or 16 RDIMM slots. With 256GB RDIMMs, the cards support up to 3.5TB or 4TB of DRAM capacity. The company said the cards are scheduled to ship to Micron in September for installation in Abaco system racks for testing, qualification, and benchmarking.

Primemas also said its Hublet architecture provides controller silicon that aggregates terabytes of DRAM for the Abaco architecture. The company also introduced SLiM, or Switchless Pooled Memory, a 1U memory architecture for AI inference workloads that need larger KV cache capacity. Primemas said SLiM lets multiple servers access tens of terabytes of shared DRAM without external CXL switches.

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.