Microsoft has joined the ranks of tech giants developing custom computing chips to manage the high costs associated with delivering artificial intelligence (AI) services. The company unveiled two new chips, Maia and Cobalt, at its Ignite developer conference in Seattle, marking a significant step in bringing key technologies in-house to optimize AI services.
Maia Chip: Designed for AI computing tasks, the Maia chip is tailored to run large language models like those underlying Microsoft's Azure OpenAI service. This chip is integral to Microsoft's $30-a-month "Copilot" service for business software users and developers aiming to create custom AI services. Maia's development, in collaboration with OpenAI, reflects Microsoft's strategy to handle the considerable expenses of AI services, which can be up to 10 times more than traditional services like search engines.
Cobalt Chip: The Cobalt chip, a central processing unit (CPU) made with technology from Arm Holdings, is part of Microsoft's broader strategy to compete with Amazon Web Services (AWS) and its in-house Graviton chips. Cobalt has been tested in powering Microsoft's business messaging tool, Teams, and the company plans to offer direct access to Cobalt to its Azure customers.
Microsoft's decision to develop these chips is a response to the growing need for specialized hardware to efficiently handle AI's computational demands. By using these chips in its Azure cloud service and subscription software offerings, Microsoft aims to provide better, faster, and more cost-effective AI solutions to its customers.
Both Maia and Cobalt are made with 5-nanometer manufacturing technology from Taiwan Semiconductor Manufacturing Co. Notably, the Maia chip will use standard Ethernet network cabling, a shift from the custom Nvidia networking technology previously used in Microsoft's supercomputers for OpenAI. This move towards standardization is likely to reduce costs and improve scalability.
As the AI industry continues to grow, Microsoft's chips represent a strategic investment in the future of AI and cloud computing. Their development not only positions Microsoft competitively in the cloud market but also demonstrates the company's commitment to innovation and efficiency in AI technologies.