Microsoft Unveils New Chips to Power AI and Cloud at Ignite
Microsoft will soon be deploying its own custom chips to power the company's AI and cloud services.
Two new chips, the Microsoft Azure Maia AI Accelerator and the Microsoft Azure Cobalt CPU, were unveiled during Wednesday's Ignite conference in Redmond, Washington. Their goal is to further power Microsoft's AI services, like Copilot and Azure OpenAI, by providing computing power optimized for AI and generative AI tasks, and increase efficiency in its cloud computing datacenters.
"Microsoft is building the infrastructure to support AI innovation, and we are reimagining every aspect of our datacenters to meet the needs of our customers," said Scott Guthrie, executive vice president of Microsoft's Cloud and AI Group. "At the scale we operate, it's important for us to optimize and integrate every layer of the infrastructure stack to maximize performance, diversify our supply chain and give customers infrastructure choice."
The two chips, which will be manufactured by the Taiwan Semiconductor Manufacturing Company (TSMC), are scheduled to roll out to Microsoft's datacenters sometime "early next year." Microsoft has no plans to offer these custom chips to organizations.
The first chip, Microsoft Azure Maia AI Accelerator, will be headed to Microsoft's Azure datacenters, and will be tasked with increasing optimization for GitHub Copilot and its OpenAI integrations, like Microsoft's Bing Chat, which received a name change to Microsoft Copilot during Ignite. According to Microsoft, the chip has been specially designed for the Azure stack, and has been tested by OpenAI to run its upcoming ChatGPT 4 Turbo upgrade.
"The architecture and implementation are designed with power efficiency in mind," said Sam Altman, CEO of OpenAI, in a Microsoft announcement. "We’re making the most efficient use of the transistors on the silicon. Multiply those efficiency gains in servers across all our datacenters, it adds up to a pretty big number."
Microsoft's Azure Cobalt CPU, will aim to bring more efficiency to its cloud services. The chip will use an Arm-based design to provide energy efficiency through lower power consumption and temperatures, in an attempt to reach the company's sustainability goals.
The two new chips mark a change for the company, which had relied on third-party vendor hardware to power its services and datacenters. In fact Microsoft's datacenters were fully compiled using outside hardware. With the introduction of the new chips, Microsoft said that the silicon powering its datacenters gets one step closer to being fully designed in house.
"Microsoft innovation is going further down in the stack with this silicon work to ensure the future of our customers' workloads on Azure, prioritizing performance, power efficiency and cost," said Pat Stemen, partner program manager on the Azure Hardware Systems & Infrastructure team "We chose this innovation intentionally so that our customers are going to get the best experience they can have with Azure today and in the future."
Further, Microsoft said that it is already working on the next generation of both the Azure Maia AI Accelerator and the Microsoft Azure Cobalt CPU.