Microsoft unveils Maia 300 AI chip to challenge Nvidia
Microsoft has launched its latest in-house AI accelerator, built to run AI inference across its Azure cloud. The company plans multi-gigawatt deployments starting in 2027.
By Standard Topic. AI-generated from the reporting listed under Coverage. Published . Updated .
This story is no longer being updated. It reflects reporting up to .
What happened
Microsoft has unveiled Azure Maia 300, the newest generation of its in-house AI accelerator chips, designed to power AI inference across Microsoft Azure. The company plans to scale its Maia accelerator family to multi-gigawatt deployments from 2027, making it the primary inference system for its own products. The chip is presented as a direct challenge to Nvidia's dominance in the AI chip market.
Why it matters
The move is a direct challenge to Nvidia's market dominance in AI chips. Microsoft will rely on Maia as the main inference system for its own products.
Timeline
- Oct 14, 9:06 PM ET: Microsoft unveiled its next-generation proprietary AI chip, challenging Nvidia's market dominance.
- Oct 15, 8:12 AM ET: Microsoft announced the launch of Azure Maia 300, targeting a multi-gigawatt buildout.
- Oct 16, 2:02 AM ET: Microsoft said it will scale Maia to multi-gigawatt deployments from 2027, drawing questions about its efficiency metric.
Coverage
- Microsoft Commits Maia to Multi-Gigawatt Inference, and Its Efficiency Metric Draws Questions (Steef-Jan Wiggers)
- Microsoft announces latest in-house chip and targets multi-gigawatt buildout (nationaltechnology.co.uk)
- Microsoft Unveils Maia 300, Challenging Nvidia’s AI Chip Dominance (Park Jung-hwan)
More from Technology & AI
- China puts humanoid robots to work serving food and sorting parcels
- Newsom proposes sweeping AI safety rules for Washington
- Google releases Gemini 4 Neon AI model with million-token context
- Microsoft and Intel widen their cloud chip partnership
- Google DeepMind rolls out Genie 4 to select subscribers
- AI helped AMD shorten chip design work, OpenAI says