Marvell Introduces Structera X Industry-First Six-Channel CXL 3.2 Memory Expander for AI Inference

Newsquawk · 2026-10-12

Marvell has unveiled the second-generation Structera X 3406 and 3506 CXL 3.2 memory expanders designed to scale capacity for AI inference workloads. Featuring six DDR4 and DDR5 channels—a 50 percent increase over the previous generation—along with PCIe Gen 6 connectivity, these expanders support up to 9TB of DDR4 or 6TB of DDR5 per controller. To address soaring memory requirements, the devices integrate multi-host sharing and inline LZ4 compression, delivering up to 2 to 3 times greater effective capacity and achieving a 40 to 50 percent faster time to first token at equivalent context lengths. Currently being demonstrated at the OCP Summit, the Structera X 3406 and 3506 are scheduled to begin sampling in the first quarter of 2027, building upon the existing Structera X 2404 and 2504 architecture currently shipping to hyperscalers.

AI 시장 분석

Marvell has unveiled Structera X, the industry's first 6-channel CXL 3.2 memory expander of the second generation designed for AI inference. This product delivers a 50% increase in channels compared to the previous generation, supports PCIe Gen 6, and alleviates memory bottlenecks in AI computing. The proliferation of related technologies is expected to drive efficiency improvements and earnings growth across the AI semiconductor ecosystem.

상승 영향

AI — Marvell's CXL 3.2 memory expander improves AI inference speed by 40-50% and expands effective memory capacity by 2-3 times, directly driving efficiency and demand for AI infrastructure.
Semiconductors — The unveiling of the next-generation memory expander supporting 6-channel DDR4 and DDR5 and PCIe Gen 6 is expected to boost revenue growth for high-performance semiconductor and data center component suppliers.

DYAX 전담 분석

Marvell's new product, Structera X, supports up to 9TB of DDR4 or 6TB of DDR5 per controller and expands effective capacity by 2 to 3 times through inline compression, improving AI inference speed by 40 to 50%. This directly translates to reduced memory costs and maximized AI performance for large-scale data centers and hyperscalers, accelerating Marvell's related revenue growth.

The key points to watch going forward are the continuous increase in AI inference demand and the adoption rate by major customers. In a bullish scenario, expanded AI infrastructure investment leads to strong earnings for Marvell, while in a conservative scenario, margin pressure may arise from competitors launching similar CXL products. Therefore, sample shipment schedules following the OCP Summit must be closely monitored.

AI가 생성한 분석으로 투자 자문이 아닙니다.

DYAX Investor Sentiment

Bullish (Long) 73% · Bearish (Short) 27%

396 participants

Related News

원문 보기 — Newsquawk