Microsoft plans huge deployment of its own AI chips, with an eye on saving costs
10 Articles
10 Articles
Microsoft Ramps Up Homegrown AI Chip Push
Microsoft is installing the latest version of its Maia chips in data centers to power its AI software using mostly homegrown hardware. The company expects to stock multiple gigawatts of its Maia chips, with the Maia 300 starting to roll out next year. Bloomberg's Matt Day joins Ed Ludlow on "Bloomberg Tech." (Source: Bloomberg)
Microsoft plans huge deployment of its own AI chips, with an eye on saving costs
Microsoft's Maia 300 is its third-generation AI chip, designed for coding and complex tasks
Microsoft today announced that starting next year, it will widely adopt its own Azure Maia AI chip, consuming several gigawatts of power. AI functionality in its products and services will primarily be handled by Maia. Microsoft's second-generation chip, Maia 200, is already in use in the US, Europe, and Australia, currently handling AI functionality for Microsoft Copilot. GitHub Copilot and the Microsoft Foundry developer platform will also beg…
Microsoft to deploy over a million Maia 300 AI chips to cut Nvidia reliance
Microsoft plans to deploy over a million of its third-generation Maia 300 AI chips next year, a direct bid to cut its reliance on Nvidia hardware. Content originally published on The Tech Portal - Global technology news, latest gadget news and breaking tech news.
Microsoft continues to screw on its own AI hardware and aims to reduce dependence on classic graphics processors. Today, the company has new details on its own Maia chips for...To post: Microsoft Azure Maia: Own AI chips are being massively expanded Where to follow us: Facebook, Reddit, Google News, X, Threads Stay up to date? Adds us to Google as the preferred source!
Microsoft stated that Maia 300 chips will perform approximately 40% more useful computing per megawatt than the "leading GPU-based systems"
Coverage Details
Bias Distribution
- 67% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium













