法国公司OVHcloud（OVH.PA）首席执行官于周二表示，该公司预计将在2027年初推出其首个人工智能模型，并补充称，该公司约有一半的连接云服务可供新客户使用。
OVHcloud (OVH.PA) expects to launch its first AI model in early 2027, its chief executive said on Tuesday, while adding that about half of the French company's connected cloud was available for new clients.
可用算力意味着这家云运营商无需进行基建支出即可招揽客户，而竞争对手则必须投资额外的算力来满足地区需求。
The available capacity means the cloud operator can onboard clients without spending on construction, while rivals must invest in additional capacity to serve regional demand.
在OVH发布年度财报后，首席执行官奥克塔夫·克拉巴（Octave Klaba）对路透社表示，该公司拥有250兆瓦的连接算力，其中140兆瓦正在运行，剩余110兆瓦已准备好承接新业务。
The company has 250 megawatts of capacity connected, of which 140 MW is in operation, leaving 110 MW ready to take on new business, CEO Octave Klaba told Reuters after OVH published its annual earnings report.
克拉巴还表示，专为OVH内部数据进行调优并运行电子邮件等办公应用的生产力平台OVHai Workspace将很快发布首个试点项目，该公司首个人工智能模型将于明年紧随其后推出。
Klaba also said OVHai Workspace, a productivity platform tuned on OVH's internal data running workplace apps such as email, would release its first pilot shortly, followed by the company's first AI model next year.
克拉巴称，该模型将采用混合专家设计，将任务分派给专门的子模型。他在6月曾告诉路透社，OVH已在欧洲运行速度最快的超级计算机“朱庇特”（Jupiter）上启动了自己的专门人工智能模型，使该公司成为有望挑战Mistral的欧洲竞争者。
The model will use a mixture-of-experts design, which routes tasks to specialised sub-models, Klaba said. He had told Reuters in June that OVH had started its own AI models on Jupiter, Europe's fastest supercomputer, positioning the company as a potential European challenger to Mistral.
克拉巴重申了他对欧洲国家资助的计算项目的批评，称其一半来自欧盟、一半来自成员国的融资方式，自然会将项目推向国家级倡议。
Klaba reiterated his criticism of Europe's state-backed computing programmes, saying their financing — half EU, half member state — naturally pushed projects toward national initiatives.
他表示，这种逻辑催生了一场国家级数据中心项目的竞赛，却没有一个项目拥有与其匹配的客户基础。
He said this logic had produced a race of national data centre projects, none backed by a customer base to match.
上个月，克拉巴曾表示，欧盟计划中的人工智能超级工厂的国家战略“注定失败”，且拟议的在五年内斥资2亿欧元采购人工智能服务的公共承诺太少，无法支撑所需的私人投资。
Last month, Klaba had said that national strategies for the European Union's planned AI gigafactories were "doomed to failure" and that a proposed €200 million public commitment to buy AI services over five years was too small to underpin the required private investment.