Marvell unveils second-generation Structera X for AI inference

Marvell: Marvell Technology designs and develops semiconductor solutions focused on data infrastructure for data centers, cloud, and AI applications. The company emphasizes open, standards-based architectures to support scalable and efficient AI infrastructure deployments. In the context of this news, Marvell is introducing its second-generation Structera X product as part of ongoing efforts to address AI data center challenges showcased at industry events like OCP Global Summit 2026. Structera X: Structera X is Marvell's line of CXL memory expansion and acceleration controllers designed for pooling and extending memory resources across AI accelerators and servers. The second-generation version introduces industry-first six-channel support for the CXL 3.2 standard specifically optimized for AI inference workloads. This positions Structera X as a key component in Marvell's broader portfolio for overcoming memory capacity limitations in modern AI systems. CXL Adoption Trend: Industry efforts around CXL 3.2 standards are advancing to enable higher-capacity and more flexible memory expansion for AI, cloud, and data center environments. Memory Wall Challenge: Marvell positions its Structera platform as a solution to the AI memory wall, where accelerator compute scales faster than per-accelerator memory capacity and bandwidth. AI Infrastructure Focus: Marvell is highlighting its comprehensive AI data center solutions, including Structera CXL memory expansion, at the OCP Global Summit 2026 to support hyperscalers seeking flexible, open architectures.