How AI Describes 25 of the Largest AI Data Center Projects
This November, Five Blocks is sharing more research than we ever have: more than 10 new studies, built with data from AIQ and IMPACT, our own platforms for tracking what AI engines and Google say about companies and their leaders. We would love your feedback. If there is research you are doing, or research you would like to do together with us, get in touch. See all #DataNovember studies.
#DataNovember, Day 3: industry spotlight. We asked eight AI engines two questions about 25 of the largest AI data center projects in the US, every day in September 2026: “Tell me about [project]” and “What challenges does [project] face?”
What we found
- Two questions, two stories. The overview reads like a press release: 0 of 1,200 rated overviews were negative. Ask about challenges and 55% turn negative.
- Engine by engine. Google AI Mode was negative in 88% of challenge answers. ChatGPT in 6%.
- Hardest hit. More than 8 in 10 challenge answers were negative for Fermi America’s Project Matador, Meta Hyperion, Stargate and Vantage Lighthouse. For AWS Project Rainier and Microsoft Fairwater Atlanta, about 1 in 5 or fewer were.
- Same company, different story. Microsoft ranges from 1 in 5 negative in Atlanta to 7 in 10 in Wisconsin.
- Until September 24, Copilot’s overview said Fairwater Wisconsin did not exist.
Download the report
Watch the video
Watch on YouTube · 5 minutes
Is your project one of the 25?
We will walk you through what AI says about it, and why.
Data: Five Blocks AIQ, September 1 to 30, 2026, eight AI engines. See all #DataNovember studies.
