Gimlet Labs

Multi-silicon inference cloud platform that allows AI workloads to run simultaneously across diverse hardware types (CPUs, GPUs, high-memory systems). Claims 3x-10x inference speedup. Partners with NVIDIA, AMD, Intel, ARM, Cerebras.