Technology

Those datacenters are real. AI companies aren't using their money to build empty buildings. They're buying enormous amounts of computer hardware off the market to fill them.

https://blogs.microsoft.com/blog/2025/09/18/inside-the-worlds-most-powerful-ai-datacenter/

Today in Wisconsin we introduced Fairwater, our newest US AI datacenter, the largest and most sophisticated AI factory we’ve built yet. In addition to our Fairwater datacenter in Wisconsin, we also have multiple identical Fairwater datacenters under construction in other locations across the US.

These AI datacenters are significant capital projects, representing tens of billions of dollars of investments and hundreds of thousands of cutting-edge AI chips, and will seamlessly connect with our global Microsoft Cloud of over 400 datacenters in 70 regions around the world. Through innovation that can enable us to link these AI datacenters in a distributed network, we multiply the efficiency and compute in an exponential way to further democratize access to AI services globally.

An AI datacenter is a unique, purpose-built facility designed specifically for AI training as well as running large-scale artificial intelligence models and applications. Microsoft’s AI datacenters power OpenAI, Microsoft AI, our Copilot capabilities and many more leading AI workloads.

The new Fairwater AI datacenter in Wisconsin stands as a remarkable feat of engineering, covering 315 acres and housing three massive buildings with a combined 1.2 million square feet under roofs. Constructing this facility required 46.6 miles of deep foundation piles, 26.5 million pounds of structural steel, 120 miles of medium-voltage underground cable and 72.6 miles of mechanical piping.

Unlike typical cloud datacenters, which are optimized to run many smaller, independent workloads such as hosting websites, email or business applications, this datacenter is built to work as one massive AI supercomputer using a single flat networking interconnecting hundreds of thousands of the latest NVIDIA GPUs. In fact, it will deliver 10X the performance of the world’s fastest supercomputer today, enabling AI training and inference workloads at a level never before seen.

Hard drives haven't been impacted nearly much as memory, which is the real bottleneck, but when just one AI company, OpenAI, rolls up and buys 40% of global memory production capacity's output, it'd be extremely unlikely that we wouldn't see memory shortages for at least a while, since it takes years to build new production capacity. And then you have other AI companies who want memory. And purchases of memory from companies who are, as a one-off, extending their PC upgrade cycle, due to the current shortage who will also be competing for supply. If you have less supply relative to demand of a product, price goes up to the new point where the available amount of memory people are willing to buy at that new price point matches what's actually available. Everyone else gets priced out. And it won't be until either demand drops (which is what people talking about a 'bubble popping' are thinking might occur, if the AI-infrastructure-building effort stops sooner than expected), or enough new production capacity comes online to provide enough supply, that that'll change. Memory manufacturers are building new factories and expanding existing ones, and we've had articles about that. But it takes years to do that.