AI基础设施公司Runware推出了一款名为Sonic Inference Pod的模块化数据中心产品[1]。该设备设计用于提供推理服务，相比其他无服务器推理平台和GPU云服务，成本更低、质量更高[1]。
Runware采用分布式计算架构构建这一产品，使其可快速扩展容量[1]。该pod配备闭环冷却系统，无需水冷却，建设周期仅需数天，相比传统数据中心的数月至数年大幅缩短[1]。CEO Flaviu Radulescu表示："分布式计算架构、靠近终端用户以实现更快推理速度，长期来看将占据优势地位"[1]。
目前Runware已在美国、欧洲和亚太地区部署了10个pod[1]，已为Higgsfield AI和Wix等公司提供推理服务[1]。公司拥有160个站点可用于部署pod[1]。12月，Runware获得了5000万美元的Series A融资[1]。
Runware, an AI infrastructure company, has introduced the Sonic Inference Pod, a modular and transportable data center unit designed to deliver inference services at lower costs and higher quality compared to competing serverless platforms and GPU cloud providers [1]. The company currently operates ten pods across the United States, Europe, and the Asia-Pacific region, with 160 additional sites available for future deployment [1].
The Sonic Inference Pod features a distributed computing architecture that enables rapid capacity scaling without requiring water cooling systems, with deployment timelines measured in days rather than the months or years typical of traditional data center construction [1]. CEO Flaviu Radulescu stated that "We believe distributed compute, positioned closer to end users for faster inference, is what will win in the long term" [1]. The company has already begun serving clients including Higgsfield AI and Wix [1].
Runware's expansion comes following its December Series A funding round, which raised $50 million [1].