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

XCENA raised $135 million in Series B at a $570 million valuation, bringing total funding to $185 million.

The startup is building the MX1 chip for near-memory AI processing, using CXL to handle preprocessing, caching, and KV cache management inside the memory module.

Mass production is planned at Samsung’s foundry in 2026, with revenue expected in 2027.

The approach could reduce AI infrastructure cost and power use by moving routine inference work closer to memory.