This chip startup just raised $135M on a bet that AI’s biggest bottleneck isn’t compute — it’s memory

TechCrunch
Startup XCENA raised $135 million to develop memory-centric chips that reduce AI infrastructure costs by processing data near the memory.

Summary

XCENA, a startup founded by former Samsung and SK Hynix veterans, has raised $135 million to address AI bottlenecks caused by inefficient data movement between memory and processors. By placing compute capabilities directly within memory modules using their MX1 chip, XCENA aims to eliminate the need for costly round trips to CPUs and GPUs. The company plans to move to mass production by 2026, targeting hyperscalers who seek to reduce infrastructure costs by shifting toward more memory-centric architectures.

(Source:TechCrunch)