Anthropic 正在打造自研芯片

Anthropic 正在组建团队设计自研 AI 芯片，以应对其模型需求激增的局面。

根据一则招聘信息和公司声明，这家 AI 创业公司正在招聘工程师，为其 Claude AI 模型设计定制芯片。

Anthropic 发言人向 Business Insider 表示，公司正在组建内部硅芯片团队，为 Claude 设计定制芯片；这是该公司首次公开确认这一计划。

该发言人补充称，Anthropic 将协同设计硬件和模型，从而使 Claude 能以“客户所需的规模”更快、更高效地运行。

该发言人表示，Anthropic 一直并将继续采取“多芯片策略”，AWS、Google、Nvidia 和 AMD 的硬件仍将是这家 AI 实验室扩展规模工作的核心。

路透社此前报道称，Anthropic 正在探索设计自有芯片的可能性；而 The Information 上个月报道，Anthropic 曾与三星电子洽谈，后者可能成为其制造合作伙伴。

近期发布的一则招聘信息详细说明了 Anthropic 的“定制硅芯片团队”计划。该公司正寻求在广泛芯片设计和验证领域具备专业能力的工程师。

该工程师职位标注的年薪范围为 32 万至 48.5 万美元，要求候选人能够证明自己曾对半导体设计的最终定稿和交付作出“直接的个人贡献”。

“这是一个面向曾将芯片设计交付落地、对项目进度有现实认知，并且能够在没有庞大组织支撑的情况下作出重要决策的人选的岗位，”招聘信息写道。

在内部设计 AI 芯片是一个成本高昂且困难的过程，但这可以帮助企业降低对主导 AI 加速器芯片市场的 Nvidia 的依赖。

这也使 AI 实验室能够定制算力，以匹配自身模型的需求。

Anthropic 是最新一家开始设计自有芯片的 AI 实验室。今年 6 月，竞争对手 OpenAI 发布了其与 Broadcom 共同开发的专用半导体，名为 Jalapeño。

根据路透社查阅的一份内部备忘录，Meta 也计划于 9 月将其下一代 AI 芯片投入生产；AI 创业公司 Mistral 的 CEO 则表示，这家法国公司正在考虑打造自己的硅芯片。

Anthropic is building a team to design its own in-house AI chips as it faces surging demand for its models.

The AI startup is hiring engineers to design custom chips for its Claude AI model, according to a job listing and a company statement.

An Anthropic spokesperson told Business Insider that the company is building an in-house silicon team to design custom chips for Claude, the first time it has publicly confirmed the plans.

They added that Anthropic would co-design hardware and models, allowing Claude to run faster and more efficiently "at the scale our customers need."

The spokesperson said Anthropic has taken and will continue to take a "multi-chip approach" in which hardware from AWS, Google, Nvidia, and AMD will remain central to the AI lab's scaling efforts.

Reuters previously reported that Anthropic was exploring the possibility of designing its own chips, while The Information reported last month that Anthropic had held talks with Samsung Electronics as a potential manufacturing partner.

A recently posted job listing detailed Anthropic's plans for a "custom silicon team," with the company seeking engineers with expertise across a wide range of chip design and verification domains.

The engineer position, which lists a salary range of $320,000 - $485,000, says candidates must be able to demonstrate "direct personal contribution" to the finalization and shipping of semiconductor designs.

"This is a role for someone who has shipped silicon, has a realistic relationship with schedules, and is comfortable making consequential calls without a large organization behind them," the listing read.

Designing an AI chip in-house is a costly and difficult process, but it can help companies to reduce their reliance on Nvidia, which dominates the market for AI accelerator chips.

It also allows AI labs to customize their compute to fit their models' needs.

Anthropic is the latest AI lab to move into designing its own chips. In June, rival OpenAI unveiled its own purpose-built semiconductor, dubbed Jalapeño, which it developed with Broadcom.

Meta is also planning to put its next-generation AI chip into production in September, according to an internal memo reviewed by Reuters, and the CEO of AI startup Mistral has said the French company is considering building its own silicon.