Gate News: On March 17, decentralized GPU computing infrastructure Aethir officially launched its managed Kubernetes service, marking its strategic expansion into the enterprise AI infrastructure market. Kubernetes (K8s), as an open-source container orchestration industry standard, can automatically manage large-scale server clusters. Aethir’s move aims to provide global AI teams with high-performance development environments that do not require managing underlying hardware. Key features of the service include: minutes-level cluster deployment via self-service API; support for the latest generation GPUs such as NVIDIA H100, H200, B200; starting at $1.45 per hour with no outbound data fees; enterprise-grade security with zero-trust access control, RBAC, and comprehensive audit logs. Aethir reduces overall computing costs by eliminating middleman premiums and virtualization expenses, lowering costs by 60% to 80%. Currently, Aethir’s decentralized infrastructure covers 95 countries and regions worldwide, with 435,000 GPU containers, providing low-latency computing access across more than 200 physical nodes, serving over 150 enterprise clients.