Report: Nvidia to acquire AI model repository Hugging Face for $13 billion
Report: Nvidia to acquire AI model repository Hugging Face for $13 billion
报道：英伟达拟以 130 亿美元收购 AI 模型库 Hugging Face
Nvidia is reportedly moving forward to acquire Hugging Face for $12.9 billion. The acquisition could help the hardware giant expand and fortify its deep integration with the wider AI industry. 据报道，英伟达（Nvidia）正推进以 129 亿美元收购 Hugging Face 的计划。此次收购将有助于这家硬件巨头进一步扩大并巩固其与整个 AI 行业的深度整合。
Among other things, Hugging Face is a cloud repository for AI models, similar in some respects to what GitHub is for conventional computer software. Developers and researchers search for models that meet certain criteria, download and run them, and fine-tune them into different variants that then get uploaded back up to Hugging Face. Hugging Face 是一个 AI 模型云存储库，在某些方面类似于传统计算机软件领域的 GitHub。开发者和研究人员可以在平台上搜索符合特定标准的模型，下载并运行它们，将其微调为不同的变体，然后再上传回 Hugging Face。
The deal is not finalized, so it could still run into trouble ahead, but both parties appear to be actively pursuing it, according to a report from The Information, which cites one person with knowledge of the matter. Another source confirmed to CNBC that talks are underway. Hugging Face reportedly had other companies trying to buy it, too, including Salesforce. 据《The Information》援引一位知情人士的消息称，该交易尚未最终敲定，因此未来仍可能面临变数，但双方似乎都在积极推进。另一位消息人士向 CNBC 证实，谈判正在进行中。据报道，此前还有包括 Salesforce 在内的其他公司曾试图收购 Hugging Face。
Nvidia had invested in Hugging Face in the past, but it wasn’t the only one—Google and Microsoft had also invested. The company was founded in 2016. 英伟达过去曾投资过 Hugging Face，但并非唯一投资者——谷歌和微软也曾参与投资。该公司成立于 2016 年。
Nvidia and several other tech companies have publicly supported the role of open-weight models as alternatives to the proprietary closed-frontier models trained by companies like OpenAI and Anthropic. Nvidia has even released its own family of open models, called Nemotron. By acquiring Hugging Face, Nvidia gains significant influence in that ecosystem. 英伟达和其他几家科技公司已公开支持“开放权重模型”（open-weight models），将其作为 OpenAI 和 Anthropic 等公司开发的专有闭源前沿模型的替代方案。英伟达甚至发布了名为 Nemotron 的自有开放模型系列。通过收购 Hugging Face，英伟达将在该生态系统中获得巨大的影响力。
Additionally, Nvidia previously had plans for its own cloud AI business but had trouble getting it off the ground. This could help get that engine running again. Hugging Face’s business is reportedly not yet profitable, but the fact that it has become the de facto place to store, download, or work on open-weight models makes it highly strategically relevant at this moment, regardless of its current bottom line. 此外，英伟达此前曾有开展自有云 AI 业务的计划，但进展并不顺利。此次收购可能有助于重启该业务。据报道，Hugging Face 目前尚未盈利，但它已成为存储、下载或开发开放权重模型的事实标准平台，这使其在当前具有极高的战略意义，而不论其目前的财务状况如何。
The acquisition is particularly important for Nvidia, as OpenAI, Anthropic, and other frontier labs have begun investing in their own specialized hardware on which to train and run those models, seeking vertical integration and freedom from Nvidia’s leverage. With Hugging Face in its portfolio, Nvidia may take measures to encourage that side of the market to continue relying on its hardware. 此次收购对英伟达尤为重要，因为 OpenAI、Anthropic 和其他前沿实验室已开始投资研发自有的专用硬件来训练和运行模型，旨在实现垂直整合并摆脱对英伟达的依赖。将 Hugging Face 纳入麾下后，英伟达可能会采取措施，鼓励该市场继续依赖其硬件。
Hugging Face is best known for hosting large language models, but it has increasingly invested in the types of models used in robotics and other physical AI applications. Nvidia is already one of the biggest players in that space, so the acquisition may open opportunities there in the future. Hugging Face 以托管大语言模型而闻名，但它也越来越多地投资于机器人和其他物理 AI 应用中使用的模型。英伟达已经是该领域最大的参与者之一，因此此次收购可能会在未来为该领域带来更多机遇。