Microsoft’s AI chips drive 40% efficiency gains, says Nadella

Microsoft just posted one of its strongest quarters in recent memory, and CEO Satya Nadella wants everyone to know the company’s homegrown AI chips are a big reason why. During the fiscal Q4 2026 earnings call, Nadella revealed that Microsoft’s custom silicon has delivered a 40% improvement in inference throughput for its Copilot models, a metric that essentially measures how quickly AI can process and respond to user requests.

Investors liked what they heard. Microsoft shares jumped more than 8% in after-hours trading following the report.

The Maia 200 factor

The star of the show is the Maia 200, Microsoft’s custom AI accelerator chip that was announced in January 2026 and is now humming along in data centers in Iowa and Arizona. According to the company, the chip delivers over 30% better tokens per dollar than the previous hardware in Microsoft’s fleet.

Microsoft’s MAI models running on the Maia 200 also saw a matching 40% performance increase.

Still, Microsoft isn’t abandoning its existing chip suppliers. The company maintains strategic partnerships with both Nvidia and AMD, adopting what amounts to a multi-vendor strategy.

Azure and AI revenue tell the bigger story

Azure revenue grew 43% year-over-year in the quarter ending June 30, 2026.

Microsoft’s AI business as a whole hit an annual recurring revenue run rate above $37 billion. That figure represents a 123% increase from the prior year.

The growth comes at a cost, though. Microsoft spent approximately $41 billion on AI infrastructure capital expenditures in the quarter.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.