微软正在推出自己的人工智能芯片,以减少对Nvidia和AMD的依赖,提高效率和控制力。 Microsoft is rolling out its own AI chips to cut reliance on Nvidia and AMD, boosting efficiency and control.

微软公司正在加速转向为数据中心定制的AI芯片,自2023年以来部署了Azure Maia和Cobalt处理器,并开发了下一代版本。 Microsoft is accelerating its shift to custom-designed AI chips for data centers, deploying its Azure Maia and Cobalt processors since 2023 and developing a next-generation version. 这一行动由CTO Kevin Scott牵头,旨在全面控制基础设施,以提高绩效和效率,减少对Nvidia和AMD的依赖。 The move, led by CTO Kevin Scott, aims to gain full control over infrastructure for better performance and efficiency, reducing reliance on Nvidia and AMD. 微软与亚马逊和字母一致, 尽管供应持续短缺, 今年在AI投资超过3000亿美元。 This aligns Microsoft with Amazon and Alphabet, all investing over $300 billion in AI this year despite ongoing supply shortages. Scott说需求仍然超过供给, 使得很难跟得上自查特格伯特启动以来的AI增长。 Scott said demand still outpaces supply, making it hard to keep up with AI growth since ChatGPT's launch.