• Microsoft is constructing a multibillion-dollar 'super factory' AI data center in Mount Pleasant, Wisconsin, designed specifically for intensive AI model training.
  • The Fairwater facility will house hundreds of thousands of Nvidia GPUs and deliver up to ten times the performance of today's fastest supercomputer when it comes online in early 2026.
  • The project represents the latest escalation in the AI infrastructure race, with Microsoft committing approximately $80 billion to AI-enabled data centers in FY2025 alone.

Microsoft is building what it describes as the world's most powerful AI data center—dubbed the Fairwater facility—in Mount Pleasant, Wisconsin, as part of a massive infrastructure expansion to support the exploding computational demands of artificial intelligence training for both its own services and OpenAI. The specialized data center, which the company internally refers to as a 'super factory' for AI, is scheduled to begin operations in early 2026.

The Wisconsin site is engineered to host hundreds of thousands of the latest Nvidia GPUs and is projected to deliver computational performance up to ten times greater than that of the fastest existing supercomputers. This leap in capability is crucial for training the increasingly large and complex foundation models that underpin generative AI systems. According to people familiar with the company's plans, the Fairwater network is being built to supply 'multi-gigawatts' of power dedicated to AI workloads.

This buildout is a central component of Microsoft's broader strategy to maintain its leadership in the cloud and AI sectors. The company is on track to invest around $80 billion in new AI-enabled data centers in its 2025 fiscal year. The Wisconsin project alone involves two data center builds representing a capital investment of up to $7.3 billion, according to state and company announcements.

The economic impact is expected to be significant. At the peak of construction, the project is forecast to create over 3,000 jobs, with up to 800 permanent high-skill positions remaining once the facility is operational. In a move to cultivate a local technical workforce, Microsoft is partnering with area colleges to develop specialized data center training programs. A company spokesperson confirmed the job creation estimates but declined to comment on specific performance metrics for the Fairwater facility.

The project underscores the intensifying competition among U.S. tech giants to secure dominance in AI infrastructure. While companies like Google, Amazon, and Meta are also undertaking massive data center expansions, Microsoft's tight integration with OpenAI and its focus on building dedicated, high-performance 'AI factories' rather than general-purpose cloud capacity is seen by analysts as a key differentiator. The race is not just about scale but about designing entire systems—from advanced cooling to large-scale fiber connectivity—specifically optimized for the unique demands of AI model training.

Efforts to reach local environmental groups for comment on the facility's energy consumption and sustainability measures were unsuccessful. The scale of the project has fueled broader debates about the environmental footprint of advanced AI and the concentration of critical technological infrastructure.