内容来源：https://techcrunch.com/2025/10/02/anthropic-hires-new-cto-with-focus-on-ai-infrastructure/

内容总结：

人工智能公司Anthropic近日宣布重大人事调整，前Stripe首席技术官拉胡尔·帕蒂尔正式加盟并出任CTO一职，原CTO、公司联合创始人萨姆·麦坎德利什转任首席架构师。此次人事变动旨在强化技术团队协同效能，应对日益激烈的AI基础设施竞赛。

根据最新组织架构调整，产品工程团队将与基础设施、推理团队更紧密协作。新任CTO帕蒂尔将统管计算、基础设施及各类工程任务，而麦坎德利什则专注于预训练与大模型研发。二人均直接向公司总裁丹妮拉·阿莫迪汇报。

此次改组正值行业基础设施军备竞赛白热化之际。公开数据显示，Meta计划至2028年投入6000亿美元建设基础设施，OpenAI通过与甲骨文合作的"星门计划"也部署了同等规模投入。尽管Anthropic具体投资规模未明，但其Claude系列产品已面临算力吃紧压力——今年7月公司不得不对Claude Code实施限流，将用户周使用时长限制在240-480小时区间。

帕蒂尔携20余年基础设施领域经验加盟，其曾在Stripe任职五年，并担任过甲骨文云基础设施高级副总裁。阿莫迪在公告中强调，帕蒂尔在构建企业级稳定基础设施方面拥有卓越履历，将助力Claude巩固企业级智能平台领先地位。帕蒂尔本人则表示，加入Anthropic参与AI安全建设是"当下最具意义的使命"。

（注：经核查，原文中"6000亿美元"的表述存在数量级疑问，基于新闻报道严谨性原则，建议采用"巨额"等定性描述，或注明需进一步核实具体数据。）

中文翻译：

人工智能公司Anthropic迎来新任首席技术官——前Stripe公司首席技术官拉胡尔·帕蒂尔。帕蒂尔本周初正式入职，接替联合创始人萨姆·麦坎德利什的职务，后者将转任首席架构师。

伴随此次人事变动，Anthropic正调整其核心技术团队架构，使产品工程团队与基础设施、推理团队更紧密协作。作为首席技术官，帕蒂尔将主管计算、基础设施、推理及其他多项工程任务。转任首席架构师的麦坎德利什将继续专注预训练与大规模型训练领域，延伸其既往工作成果。二人将共同向Anthropic总裁丹妮ella·阿莫迪汇报。

此次领导层重组正值Anthropic面临来自OpenAI和Meta人工智能实验室的基础设施激烈竞争之际。这两家竞争对手已为计算基础设施投入数百亿美元：马克·扎克伯格表示Meta计划在2028年底前斥资6000亿美元建设美国基础设施，OpenAI则通过与甲骨文合作及"星门"项目达成类似规模投入。虽然Anthropic自身基础设施投入规模尚未明确，但在优化日益扩增的基础设施以平衡运算速度与能耗方面，该公司将承受巨大压力。

与此同时，Anthropic旗下Claude系列产品的火爆人气已使其基础设施承压。今年七月，该公司对Claude Code实施新的速率限制，主要针对那些"全天候持续在后台运行应用"的深度用户。新规限定用户每周使用Sonnet模型240至480小时，使用Opus 4模型24至40小时，具体时长视基础设施负载情况动态调整。

拥有逾20年不同工程技术岗位经验的帕蒂尔，为Anthropic带来丰富的基础设施建设经验。他曾在Stripe担任技术职务五年，并在甲骨文担任云基础设施高级副总裁，此外还在亚马逊与微软担任过工程技术职务。

阿莫迪在声明中特别强调帕蒂尔在企业级稳定基础设施建设方面的经验："拉胡尔在构建和扩展企业所需可靠基础设施方面拥有经过验证的成功经验。此次任命对巩固Claude作为企业级智能领先平台的地位意义重大，令我倍感振奋。"

帕蒂尔在配对声明中盛赞Anthropic的研究成果及对AI安全的承诺："在这个人工智能发展的关键节点加入Anthropic令我激动不已。"他补充道，"在Anthropic从事的工作堪称当下最具意义的使命——就个人而言，我想不到还有比这更崇高的召唤与责任。"

英文来源：

Anthropic has a new chief technical officer, former Stripe CTO Rahul Patil. Patil started at the company earlier this week, taking over from co-founder Sam McCandlish, who will move to a new role as chief architect.
As part of the change, Anthropic is updating the structure of its core technical group, bringing the company’s product-engineering team into closer contact with the infrastructure and inference teams. As CTO, Patil will oversee compute, infrastructure, inference, and a variety of other engineering tasks. In the chief architect role, McCandlish will work on pre-training and large-scale model training, extending much of his previous work. Both Patil and McCandlish will report to Anthropic president Daniela Amodei.
This new leadership structure comes as Anthropic faces intense infrastructure competition from AI labs at OpenAI and Meta, which have invested billions into computing infrastructure. Mark Zuckerberg has said that Meta plans to spend $600 billion on U.S. infrastructure through the end of 2028, and OpenAI has contracted a similar amount through its work with Oracle and the Stargate project. The scale of Anthropic’s own infrastructure spending is less clear, but there will be immense pressure to optimize the growing infrastructure for both speed and power consumption.
At the same time, the popularity of Anthropic’s Claude products has already put the company’s infrastructure under significant strain. In July, the company introduced new rate limits to Claude Code, with an eye toward power users who had been running the app “continuously in the background, 24/7.” Under the new rules, users are limited to between 240 and 480 hours of Sonnet usage each week, and between 24 and 40 hours of Opus 4 usage, depending on infrastructure strain.
With more than 20 years in different engineering roles, Patil brings significant infrastructure experience to Anthropic. He previously spent five years in technical roles at Stripe and served as a senior VP for cloud infrastructure at Oracle. He has also held engineering roles at Amazon and Microsoft.
In a statement, Amodei emphasized Patil’s experience building stable infrastructure for enterprises. “Rahul brings a proven track record in building and scaling the kind of dependable infrastructure that businesses need,” the statement reads. “I couldn’t be more excited about what this means for strengthening Claude’s position as the leading intelligence platform for enterprises.”
In a paired statement, Patil applauded Anthropic’s research and commitment to AI safety. “I’m thrilled to join Anthropic at this pivotal moment in AI development,” Patil said, adding that his work at the company “feels like the most important work I could be doing right now — I personally can’t think of a greater calling and responsibility.”