Dr. Gleb Tsipursky – The Office Whisperer
Nirit Cohen – WorkFutures
Angela Howard – Culture Expert
Drew Jones – Design & Innovation
Jonathan Price – CRE & Flex Expert
As demand for artificial intelligence infrastructure surges, developers in the Midwest are increasingly turning to an unexpected source for data center space: underused and mostly vacant office buildings.
Once symbols of economic strength in urban cores, many of these properties are now being repurposed to support the growing need for AI inference — the process of running AI models in real-world applications.
Unlike massive hyperscale data centers used to train AI models (typically built in rural or suburban areas with hundreds of megawatts of power) inference centers are smaller, often ranging between 20 and 60 megawatts. This makes them a natural fit for existing office towers in cities like Chicago and Minneapolis, where electricity infrastructure is already in place but underutilized due to the long-term decline in office occupancy, according to BisNow.
In Minneapolis, one developer successfully converted a largely empty office tower into an AI-focused data center, reusing the building’s electrical capacity and injecting new economic activity into the urban core.
A similar project is now underway in Chicago, where a 20 MW AI data center — marketed as the city’s first of its kind — is set to break ground by the end of the year.
Developers are increasingly viewing these downtown sites as prime opportunities to deliver edge data centers quickly, especially as businesses seek AI services closer to urban centers and end users. Office conversions often require fewer regulatory hurdles compared to new builds, and in many cases, local governments are actively supporting these transformations to offset lost tax revenue from commercial real estate vacancies.
With hybrid work reducing demand for commercial space, many office buildings have increasingly become financial liabilities. Repurposing them into data centers offers a viable path to revitalization, turning stranded power and vacant floors into revenue-generating, tech-forward infrastructure.
As AI applications proliferate across industries, the need for local processing power is expected to rise dramatically. The Midwest’s urban office buildings — many with built-in power access and connectivity — are emerging as a practical and cost-effective solution to meet this demand.
