Microsoft just flipped the switch on what it calls the world’s most powerful AI supercomputer. The stock market responded with a collective shrug.
The Fairwater AI datacenter in Mount Pleasant, Wisconsin, became fully operational by April 2026, with construction completion formally announced on June 23. Equipped with hundreds of thousands of NVIDIA GB200 GPUs, the facility is designed to supercharge Microsoft’s Azure cloud and AI ambitions. And yet MSFT shares have been stuck in a trading range between $350 and $390 for much of 2026, hovering around $383 as of July 8.
The capex problem Wall Street can’t ignore
Microsoft is projected to spend roughly $190 billion in capital expenditures during calendar year 2026. Just one day before the Fairwater completion announcement, Microsoft revealed plans for a new 2 gigawatt datacenter campus in Pecos, Texas, aimed at further expanding global capacity. The Fairwater facility is part of what Microsoft describes as a growing network of “AI superfactories” that also includes sites in Atlanta.
The company’s AI business has reached a $37 billion annualized revenue run rate, which sounds impressive until you stack it against the infrastructure investment required to sustain it.
Where crypto miners fit into the picture
Microsoft struck a deal with IREN, a former Bitcoin miner, valued at $9.7 billion. The arrangement underscores IREN’s pivot from cryptocurrency mining to AI workloads, and it’s not an isolated case. Multiple public Bitcoin miners have been securing contracts with hyperscale datacenter operators, recognizing that their core competency, managing enormous amounts of computing power and energy, translates directly to AI infrastructure.
Microsoft also launched a “Community-First AI Infrastructure” initiative back in January 2026, committing to cover full power costs and replenish water used in its datacenters.
What this means for investors
For Microsoft shareholders specifically, the stock’s sideways action reflects a market that’s pricing in execution risk rather than growth potential. The $37 billion AI revenue run rate proves demand exists. The question is whether $190 billion in annual capex eventually generates returns that justify the spend.