[Marvell to Launch Storage and Memory Solutions for AI Infrastructure, Betting on CXL and Optical Interconnect to Expand AI Infrastructure] TechFlow News, August 3 — Data infrastructure semiconductor manufacturer Marvell Technology announced that it will launch its memory and storage solutions for AI infrastructure from August 4 to 6. Marvell stated that as agentic AI inference scales up, memory capacity, bandwidth, and data connectivity are becoming as critical as computational power. Larger-scale models, longer context windows, and the growing demand for KV Cache are driving unprecedented demand for memory resources in AI systems. The company claims that by bringing data closer to compute units and reducing latency, its AI memory and storage technologies can help customers improve hardware utilization, enhance token processing efficiency, and achieve a better balance between performance, power consumption, and cost. As AI model sizes continue to grow, the traditional 'compute-first' architecture is gradually shifting toward 'compute, memory, and interconnect co-optimization.' Emerging infrastructure technologies, including CXL, high-speed storage, and optical interconnect, are expected to become key directions for the next phase of AI data center expansion.