内容来源：https://www.theverge.com/ai-artificial-intelligence/880513/nvidia-meta-ai-grace-vera-chips

内容总结：

【科技巨头Meta与英伟达达成重磅合作 斥资采购数百万AI芯片加速数据中心布局】

近日，Meta与英伟达签署了一项多年期合作协议，将为其数据中心部署数百万颗英伟达Grace CPU、Blackwell及Rubin GPU，并计划于2027年引入下一代Vera CPU。值得注意的是，此次合作标志着英伟达首次大规模独立部署其Grace架构CPU。英伟达表示，该方案将显著提升Meta数据中心的“能效表现”。

尽管Meta正自主研发AI芯片，但据《金融时报》报道，其自研芯片计划面临“技术挑战与推广延迟”。与此同时，英伟达也需应对AI基建融资中的折旧风险、芯片贷款压力及市场竞争加剧等问题。此前，因传出Meta考虑采用谷歌Tensor芯片的消息，英伟达股价单日下跌4%；而AMD近期已与OpenAI、Oracle达成芯片合作，进一步加剧行业竞争。

尽管双方未披露具体交易金额，但业内估算，Meta与微软、谷歌、亚马逊等科技巨头今年的AI投入总额，已超过历史上整个阿波罗登月计划的成本规模。此次合作凸显了全球科技企业在AI基础设施领域的激烈竞逐与持续加码。

中文翻译：

Meta已与英伟达达成一项多年协议，将为其数据中心增配数百万台英伟达Grace与Vera系列CPU，以及Blackwell和Rubin系列GPU。尽管Meta长期以来一直使用英伟达硬件支持其人工智能产品，但此次协议"标志着英伟达Grace芯片首次实现大规模独立部署"。英伟达表示，这将为Meta数据中心带来"显著的每瓦性能提升"。该协议还计划于2027年在Meta数据中心部署英伟达下一代Vera系列CPU。

Meta与英伟达的新协议锁定了数百万颗人工智能芯片
英伟达首次开始独立销售人工智能专用CPU。

Meta同时也在自主研发用于运行人工智能模型的芯片，但据《金融时报》报道，其芯片战略正面临"技术挑战与推广延迟"的困境。英伟达则需应对人工智能建设融资相关的资产贬值风险与芯片贷款问题，以及日益激烈的竞争压力。CNBC指出，去年11月传出Meta考虑采用谷歌Tensor芯片处理人工智能任务的消息后，英伟达股价应声下跌4%；而去年底AMD也相继宣布与OpenAI及甲骨文达成芯片合作协议。

虽然双方未披露具体交易金额，但Meta、微软、谷歌和亚马逊今年在人工智能领域的投入预计将超过整个阿波罗登月计划的总开支。

热门资讯速览

• 苹果将于3月4日发布重要消息
• 福特正突破物理极限研发平价电动汽车
• 折叠手机能否取代笔记本电脑？
• 因内存供应危机，Valve蒸汽平台OLED版掌机将间歇性缺货
• 斯蒂芬·科尔伯特称哥伦比亚广播公司禁播其与詹姆斯·塔拉里科的访谈内容

英文来源：

Meta has struck a multiyear deal to expand its data centers with millions of Nvidia’s Grace and Vera CPUs and Blackwell and Rubin GPUs. While Meta has long been using Nvidia’s hardware for its AI products, this deal “represents the first large-scale Nvidia Grace-only deployment,” which Nvidia says will deliver “significant performance-per-watt improvements in [Meta’s] data centers.” The deal also includes plans to add Nvidia’s next-generation Vera CPUs to Meta’s data centers in 2027.
Meta’s new deal with Nvidia buys up millions of AI chips
Nvidia’s starting to sell AI CPUs for use by themselves for the first time.
Nvidia’s starting to sell AI CPUs for use by themselves for the first time.
Meta is also working on its own in-house chips for running AI models, but according to the Financial Times, it has run into “technical challenges and rollout delays” with its chip strategy. Nvidia is also dealing with concerns about depreciation and chip-back loans used to finance AI buildouts, as well as the pressure of competition. CNBC notes that Nvidia’s stock dropped four percent after a November report about Meta considering using Google’s Tensor chips for AI, and late last year, AMD announced chip arrangements with both OpenAI and Oracle.
Nvidia and Meta did not disclose how much the deal cost, but this year’s AI spending from Meta, Microsoft, Google, and Amazon is estimated to cost more than the entire Apollo space program.
Most Popular

• Apple’s doing something on March 4th
• Ford is fighting against physics to build affordable EVs
• Who needs a laptop when you have a folding phone?
• Valve’s Steam Deck OLED will be ‘intermittently’ out of stock because of the RAM crisis
• Stephen Colbert says CBS banned him from airing this James Talarico interview