OpenAI and Oracle are building another massive AI data center.
The companies, along with Related Digital, plan to establish a large-scale campus in Saline Township, Michigan. The project will deliver more than 1 gigawatt of compute capacity - enough to power enormous AI workloads.
Construction is scheduled to begin in early 2026, pending approval from the Michigan Public Service Commission. The project is part of OpenAI and Oracle's previously announced commitment to provide 4.5GW of additional Stargate infrastructure.
The proposed campus will span 250 acres with three single-story buildings. Each building will cover 550,000 square feet.
This is a multi-billion dollar development funded by private investors and financial institutions. Power will come from DTE Energy using existing sources, supplemented by a new battery storage system that the project will finance entirely.
The campus will use a closed-loop cooling system designed to maintain water usage comparable to standard office facilities - an important detail given concerns about data centers' water consumption.
During construction, the project expects to generate over 2,500 union construction jobs. Once operational, employment projections include:
More than 450 onsite roles through Related Digital and Oracle. Around 1,500 positions across Washtenaw County supporting related operations. Thousands more indirect jobs throughout Michigan and nationwide.
"This historic, multi-billion-dollar investment will ensure that Michigan plays a leading role in developing the digital infrastructure American companies need," said Jeff T. Blau, CEO of Related Companies and chairman of Related Digital.
This Michigan campus represents the latest salvo in the AI infrastructure buildout we've been covering.
Tech giants are racing to build massive data centers to power AI development. OpenAI and Oracle's Stargate project aims to deliver 4.5GW of capacity across the US. For context, Meta just announced a $27 billion data center deal in Louisiana.
The investment reflects how capital-intensive AI has become. Training and running large language models requires enormous computing power, which requires massive physical infrastructure.
The campus has been nicknamed "The Barn" after a historic red barn that will remain at the site's entrance on Michigan Avenue. The build plan includes LEED certification for environmental standards, plus minimum 75-foot setbacks from roads and visual screening features.
These design elements suggest awareness of community concerns about large industrial facilities. Data centers can be controversial neighbors due to power consumption, noise, and visual impact.
Another multi-billion dollar AI data center shows how aggressively tech companies are building infrastructure.
OpenAI and Oracle committing to 4.5GW of Stargate capacity represents one of the largest AI infrastructure buildouts announced. This Michigan campus delivering over 1GW is a significant piece of that puzzle.
The 2,500 construction jobs and 450+ permanent roles make this economically significant for Michigan. States are competing aggressively to attract these facilities because of the employment and tax revenue they generate.
For the AI industry, the continued massive infrastructure investment signals confidence that demand will justify the expense. You don't commit billions to data centers unless you expect sustained need for computing capacity.
The closed-loop cooling system and existing power sources address some environmental concerns, though 1GW of power consumption is still substantial. The project financing its own battery storage is a positive touch.
Construction starting early 2026 means this capacity comes online in 2027 or 2028. That timeline shows how long it takes to build these facilities - even with massive budgets and urgency, you can't rush a 250-acre campus.
Michigan scoring this project reflects how the AI infrastructure boom is spreading beyond traditional tech hubs. Rural Michigan isn't Silicon Valley, but it offers land, power access, and state government support.
For investors watching AI infrastructure spending, this is more evidence the buildout is real and accelerating. OpenAI, Oracle, Meta, Google, Amazon, and Microsoft are all deploying tens of billions into data centers.
Whether that capacity gets fully utilized depends on AI adoption living up to the hype. But right now, tech companies are betting big that demand will be there.
