Microsoft Is Scaling Up Maia AI Chips to Challenge Nvidia

Microsoft is accelerating its push into custom AI chips to reduce its dependence on Nvidia. Starting in 2027, the company plans to deploy its Maia AI chips at gigawatt scale across its data centers, using them to power AI inference and generate responses across its products and services.

Microsoft CEO Satya Nadella has outlined plans to expand the use of the Maia chip family, including the Maia 200 and next-generation Maia 300. Semiconductor company Marvell Technology is helping Microsoft design and develop these custom AI accelerators, potentially creating a major business opportunity for the chipmaker.

In this TechWorld report, discover:

• Microsoft's plan to expand Maia AI chip deployment from 2027
• How Maia 200 and Maia 300 fit into Microsoft's AI infrastructure
• Why Microsoft wants to reduce its dependence on Nvidia GPUs
• Marvell Technology's role in developing custom AI accelerators
• Microsoft's reported $175 billion data center spending plan
• Why Google and other technology companies are building their own AI chips
• The trade-off between specialized AI accelerators and the flexibility of GPUs
• Marvell's ambitions in the custom AI chip and networking market