SK Hynix unveils MARS rack architecture with CXL memory pooling

SK Hynix presented its Memory-Centric AI Rack System (MARS) at the Open Compute Project Global Summit. The design pulls memory out of individual servers and pools it at the rack level over Compute Express Link fabrics, shifting AI data center design from processor-centric to memory-centric.

• MARS pools memory at the rack level instead of per server
• Servers reach the shared memory pool over CXL fabrics
• The architecture was unveiled at the OCP Global Summit