Microsoft has unveiled a new AI chip, a step the company calls its "biggest bet yet" on its own silicon and a direct move to reduce its reliance on Nvidia.

The announcement, reported by GeekWire, signals an escalation in Microsoft's strategy of designing processors in-house rather than depending on third-party suppliers for the accelerators that power its AI workloads.

Why is Microsoft building its own AI chips?

Nvidia currently dominates the market for AI accelerators, and hyperscale cloud operators pay a premium for that hardware. By designing its own chips, Microsoft gains more control over:

• Supply availability for its data-center buildouts
• Cost per unit of AI compute
• Hardware-software integration with its own cloud and AI services

Microsoft's use of the phrase "biggest bet yet" frames the new chip as its most significant silicon investment to date, exceeding prior in-house development efforts.

What does the new chip mean for the Nvidia relationship?

The stated goal is to reduce reliance on Nvidia, not eliminate it. Microsoft remains a major buyer of Nvidia GPUs, and the new chip gives the company an alternative supplier inside its own walls.

For Nvidia, each hyperscaler that designs its own AI silicon represents incremental pressure on a business that has so far benefited from concentrated demand. For Microsoft, the chip is a hedge against both pricing power and supply constraints.

Who else is pursuing this path?

Microsoft is not alone among cloud providers in bringing chip design in-house. Rival hyperscalers have followed the same logic: custom silicon tuned to a company's own workloads can outperform general-purpose accelerators on cost and efficiency at scale.

The strategy mirrors a broader industry shift in which cloud operators act as chip designers and contract manufacturers produce the resulting designs.

What happens next?

The new chip now has to prove itself in production workloads. Key open questions include:

• Deployment scale across Microsoft's data centers
• Performance against current Nvidia accelerators
• Whether the savings justify the design investment

Microsoft has framed the chip as its largest silicon commitment so far. The market will judge the bet on deployment numbers and cost outcomes, not on the label.