- Microsoft to expand its NPU to all Surface devices.
- The company is also working on an NPU (Athena) for its datacenters.
- The new chipset will improve AI capabilities across products.
- It will also help to reduce costs and save time.
Microsoft reportedly plans to bring its AI Neural Processing Unit (NPU) to all its Surface devices. According to a report from The Information (via Thurrott), the company is speeding up the work to expand its AI hardware chipsets for its Surface devices as well as for its datacenters.
According to sources, Microsoft plans to include its new NPU chips in the next generation of Surface devices in an effort to speed up AI and ML (machine learning) capabilities, which is a critical component to improve performance and new AI features, such as task automation.
Currently, the Surface Pro 9 with the Qualcomm processor is the only Surface that features a Neural Processing Unit. However, the company is also working with Intel and AMD to integrate the NPU into their processors.
On the datacenter side, since 2019, Microsoft has been working on its own NPU known as Athena, a new processor that will significantly improve the use of large-language models across its products, including Bing, GitHub, and Microsoft 365, saving time and money while eliminating the dependency on third-party AI hardware.
The company has more than 300 individuals working on Athena, with a small group of people from Microsoft and OpenAI already testing the chipset. Microsoft expects to have a final version of the AI hardware at some point in 2024, but it’s unclear when it will be available for Azure customers.