Globally distributed AI inference network that aggregates idle GPUs to deliver low-cost, high-throughput LLM serving with batch and real-time APIs. Pay-as-you-go intelligence at scale.