Search results for Compute Express Link
AI & Enterprise
CXL technology gains traction amid memory supply crunch as Korean firms step up
Memory chip supply continues to lag demand as AI spreads, pushing big tech to look for alternatives. Compute Express Link, or CXL, is drawing attention by enabling servers in the same data centre to share memory. The Information reported Google has begun deploying CXL and installing controllers to manage traffic to external memory pools. Adoption remains limited, but Nvidia plans CXL 3.1 support and South Korean firms are advancing initiatives.
AI & Enterprise
SKT, Fanessia to redesign AI data center architecture to fix inefficiency
SK Telecom said it signed an agreement with Fanessia at MWC 2026 in Barcelona to jointly develop a CXL-based next-generation AI data centre architecture. The partners plan to expand resource interconnection from inside servers to rack level and allow flexible combinations of CPU, GPU and memory. They aim to lift AI processing efficiency without unnecessary equipment expansion and to improve data centre economics.
Industry
AsicLand adds contract with Primemas for CXL 3.2 chiplet development
AsicLand said on Monday it expanded its contract with Primemas to develop a next-generation chiplet system-on-chip to 9.5 billion won. It extended the contract period to enhance technical specifications and ensure product completeness. The Falcon-1 chipset will include CXL 3.2 and a chiplet interface, and is part of Primemas\' chiplet-based SoC platform Hublet. AsicLand will handle backend design, testing and tape-out work.
