Penguin Solutions says traditional server architecture is hitting limits as AI workloads run into memory bottlenecks and data-transfer latency. It points to CXL

Penguin Solutions says traditional server architecture is hitting limits as AI workloads run into memory bottlenecks and data-transfer latency. It points to CXL as a key fix for scaling inference more efficiently; the stock recently rose 10.4%.