Microsoft launches custom Maia 300 silicon to power internal workloads

Microsoft announced a major expansion of its in-house artificial intelligence hardware program. The tech giant confirmed that its custom Maia 200 chips are currently running key internal services such as Microsoft Copilot and Teams, while its next-generation Maia 300 processors are scheduled to launch in 2027. According to a blog post by cloud executive Scott Guthrie, Microsoft intends to transition its cloud infrastructure toward multi-gigawatt custom deployments that handle the heavy computational load of serving AI applications.

For everyday consumers, custom chips help software companies run complex digital assistants and automated tools without incurring unsustainable energy costs or suffering from service slowdowns. Tasks like analyzing long documents, drafting software code, and powering multi-step automated assistants require massive processing power. By designing specialized silicon focused on power efficiency, Microsoft aims to make its everyday digital products faster, more reliable, and cheaper to operate.

Despite these technical announcements, Microsoft has not clarified how quickly this transition will diminish its financial reliance on external hardware suppliers. It also remains unclear whether outside enterprise customers will ever get direct access to rent these custom chips or if they will remain reserved for internal products. As major tech platforms build specialized internal chips, the software market could become increasingly locked into proprietary cloud environments, making cross-platform mobility far more complicated for software developers.