通用计算公司将于2026年5月15日推出用于人工智能代理的推断云平台. General Compute launches inference cloud platform for AI agents on May 15, 2026.

通用计算公司已经为人工智能代理推出了一个推断云平台, 预计将于2026年5月15日全面开放. General Compute has launched an inference cloud platform for AI agents, set for general availability on May 15, 2026. 通过使用定制的人工智能加速器和独特的架构来分离推理阶段, 该系统可优化高量大语言模型任务的性能. Using custom AI accelerators and a unique architecture that separates inference stages, the system optimizes performance for high-volume large language model tasks. 功能包括行业标准的API,对开源和自定义模型的支持以及由水电提供动力的基础设施,其功率密度低于传统GPU. Features include an industry-standard API, support for open-source and custom models, and infrastructure powered by hydroelectric energy with lower power densities than traditional GPUs.