Microsoft is ramping up its custom silicon strategy by preparing to unveil the Maia 300 AI chip this fall while securing TSMC manufacturing capacity to reduce reliance on expensive Nvidia hardware.

  • Microsoft is negotiating with TSMC to secure manufacturing capacity for over 300,000 Maia 300 chips delivered in 2027. 
  • The Maia 200 is 30% to 40% cheaper to operate than top-tier Nvidia chips for running OpenAI and Microsoft models.
  • Microsoft is attempting to woo high-profile external customers, including ongoing discussions with Anthropic. 

Microsoft (MSFT) share price jumped 2% on Monday to its highest level in over nine months as it prepares a push in its AI silicon strategy by boosting production ahead of the release of the Maia 300 chip.

According to a report by The Information, the tech giant is seeking to drastically scale up its in-house hardware capabilities to counter rising infrastructure costs and diminish its heavy dependence on Nvidia Corp (NVDA). 

MSFT’s Manufacturing Ambitions

To support the rollout, Microsoft has been negotiating with Taiwan Semiconductor Manufacturing Co. (TSMC) to secure capacity for more than 300,000 Maia 300 units targeted for delivery in 2027. The source noted that Microsoft's ultimate goal is to obtain production capacity for over 1 million units, though supply constraints for key components and ongoing capacity negotiations with TSMC could affect total volume.

The projected production target marks a steep increase from the tens of thousands of Maia 200 chips produced to date. Past efforts encountered delays after early tests failed to hit internal performance targets. As a result, deployment of the current Maia 200 line has remained limited to just two data centers located in the U.S.

In a statement to The Information regarding the production targets, Andrew Wall, Microsoft’s general manager for Azure Maia, declined to confirm specific volume figures but emphasized the long-term vision. "We expect our Azure Maia deployments to support AI workload demand measured in gigawatts," Wall stated, pointing to infrastructure scale that would eventually require millions of AI processors.

Cutting Costs and Catching Rivals

CEO Satya Nadella has long aimed to reduce the software company's reliance on Nvidia's dominant market position. Unsealed court documents from a 2022 internal email revealed Nadella lamenting that Microsoft was "a very thin layer on top of Nvidia" while absorbing high operating expenses to host AI models.

The economic rationale for the transition is central to Microsoft's strategy. Microsoft informed investors last month that running OpenAI and internal AI models on Maia 200 processors reduces operational costs by 30% to 40% compared to top-of-the-line Nvidia hardware. 

Microsoft currently lags behind hyperscale peers such as Amazon.com Inc. and Alphabet Inc.’s Google in third-party adoption of custom hardware. Google plans to manufacture over 3 million Tensor Processing Units (TPUs) this year, while Amazon's Trainium processors have secured large-scale commitments from major developers, including Anthropic.

To narrow that gap, Microsoft has been holding discussions with Anthropic to adopt Maia chips in the future.

MSFT Stock: Retail View 

Retail sentiment on Stocktwits was ‘neutral’ with ‘low’ message volumes. Retail chatter on the stock has gained 34% from the previous session. 

MSFT stock has gained 5% year-to-date. 

For updates and corrections, email newsroom[at]stocktwits[dot]com.<