| Mobile Web

Search results for CXL

AI & Enterprise

Xcena joins OCP Global Summit 2026, showcases memory infrastructure for AI inference

Fabless semiconductor company Xcena, which develops next-generation memory infrastructure, took part in OCP Global Summit 2026 held in San Jose, California from Oct. 12 to 15, it said on Thursday. The company showcased memory infrastructure technology aimed at efficient AI inference. It presented two AI inference demos using its CXL-based memory solution MX1, including sharing KV cache across servers via a CXL memory pool and accelerating vector search using near-data processing.

AI & Enterprise

South Korean AI chip firms look to expand \'sovereign AI\' on AMD open ecosystem after Lisa Su meeting

AMD Chief Executive Lisa Su met South Korean AI chip companies to expand cooperation based on an open AI computing ecosystem. South Korea’s science ministry and AMD held a roundtable in Seoul to discuss building a heterogeneous computing ecosystem and combining AMD CPUs and GPUs with locally made AI chips. Officials and executives said demand is shifting toward inference, increasing interest in mixing GPUs and NPUs. The government plans large-scale reference projects to validate full-stack systems.

AI & Enterprise

KAIST team proposes CXL-based AI data centre architecture with Meta

Researchers at the Korea Advanced Institute of Science and Technology (KAIST) and faculty startup Panmnesia, working with Meta, proposed an AI data centre architecture based on Compute Express Link (CXL). The structure links CPUs, AI accelerators and memory and extends that connectivity across an entire data centre. The team said it could connect up to 960 accelerators in one domain, about 13 times a current NVLink-based rack, and reduce data-movement latency from microseconds to hundreds of nanoseconds.