The tech major plans to roll out Microsoft Azure Maia AI Accelerator and Microsoft Azure Cobalt CPU early next year to the company’s datacentres to initially support the company's services such as Microsoft Copilot or Azure OpenAI Service
Microsoft unveils the Microsoft Azure Maia AI Accelerator and Microsoft Azure Cobalt CPU chips to address the AI demand. (Credit: Microsoft)
Microsoft has introduced two custom-designed chips, dubbed Microsoft Azure Maia AI Accelerator and Microsoft Azure Cobalt CPU, to meet the demand for artificial intelligence (AI).
The Maia AI Accelerator is streamlined for AI tasks and generative AI, while the Cobalt CPU is an Arm-based processor customised for running general purpose compute workloads on the Microsoft Cloud.
According to the American technology company, its new chips will deliver infrastructure systems that include silicon choices, software, servers, racks, cooling systems, and others.
The custom chips can be optimised with internal and customer workloads.
Microsoft plans to roll out the custom-designed chips early next year to the company’s datacentres. They will initially support Microsoft’s services such as Microsoft Copilot or Azure OpenAI Service.
Microsoft cloud + AI group executive vice president Scott Guthrie said: “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.
“At the scale we operate, it’s important for us to optimise and integrate every layer of the infrastructure stack to maximise performance, diversify our supply chain and give customers infrastructure choice.”
The silicon architecture of the chips will allow Microsoft to improve cooling efficiency along with optimising the use of its current datacentre assets.
Besides, it will help to maximise server capacity within the company’s existing footprint.
Microsoft has also announced the general availability of Azure Boost. It is a system that makes storage and networking faster by moving the processes off the host servers onto purpose-built hardware and software.
Besides, the technology company launched a preview of the new NC H100 v5 virtual machine series built for NVIDIA H100 Tensor Core GPUs.
Microsoft also plans to add the latest NVIDIA H200 Tensor Core GPU to its fleet in 2024 to support larger model inferencing with no increase in latency.