Wafer

Wafer is an AI infrastructure company that makes model inference faster and more cost-efficient. Its autonomous performance-engineering agents optimize GPU kernels and the broader serving stack—including batching, scheduling, and memory layout—for open-source large language models. Wafer offers serverless and dedicated inference endpoints, routing workloads across NVIDIA, AMD, TPUs, and other silicon to provide low-latency, high-throughput APIs for enterprise and AI-native applications.