CEA-Leti Scales Ferroelectric RAM to 22nm Node, Unlocking Denser, More Efficient Memory for Edge AI
AI-summarised brief · reviewed before publication
CEA-Leti has made a significant breakthrough in memory technology by scaling ferroelectric RAM (FeRAM) to the 22nm manufacturing node using a 3D capacitor architecture. This achievement removes a longstanding density barrier, enabling faster and more energy-efficient artificial intelligence (AI) at the edge. The team achieved memory cells that are 2.5 times smaller than standard SRAM at the same node, matching the density of SRAM at the 10nm node. FeRAM retains data without power, combining non-volatility with high density.
💡 Why It Matters
- · The widespread adoption of FeRAM could significantly reduce global electricity consumption by enabling devices to process data locally, thereby minimizing the need for cloud-based AI processing.
- · This technology paves the way for highly energy-efficient systems, helping to reduce reliance on fossil-based energy sources.