Density AI

Density AI is building a new generation of AI accelerators around a novel memory architecture intended specifically for frontier-scale LLM inference. The accelerator combines the near-memory compute topology of SRAM designs with the density of HBM, targeting energy and speed for large-model, long-context workloads while supporting larger, more reliable models than SRAM-based designs. The company is designing a full-stack solution from silicon to rack to software, integrating memory directly into the compute stack so bandwidth, capacity, and compute live in the same place. A single package tiles multiple compute dies with memory and compute replicated across a large surface so capacity, bandwidth, and FLOPs scale together.