Google takes Crusoe's Wyoming site and will spend $17B finishing it
Even a $30.9 billion neocloud can lose a project to its would-be customer. Crusoe confirmed Friday that it sold its unfinished data center project in southeastern Wyoming to Google, TechCrunch reported. Google didn't disclose the price, but announced it will invest $17 billion to develop the campus.
The backstory is the interesting part. Crusoe announced the site last year with Blackstone-backed Tallgrass, planning to scale it to 10 gigawatts. It tried and failed to land Google as the customer. Google reportedly raised concerns about cost and construction timeline under Crusoe's management, Bloomberg first reported the handover in June, and now Google is simply building it itself. The two still work together: Crusoe is developing another campus for Google near Amarillo, Texas.
None of this stopped Crusoe's momentum. It raised $3.9 billion in September, builds the Abilene, Texas campus for Oracle and OpenAI, is building a 900-megawatt campus next door for Microsoft due mid-2027, counts Meta as a customer, and per Bloomberg signed a roughly $13 billion, five-year deal with Jane Street.
Why agent people should care: inference for agents runs on exactly this kind of capacity, and the bottleneck is increasingly delivery, not demand. Power, equipment and local opposition decide when compute comes online. A hyperscaler that thinks a developer is too slow will just take the keys, which tells you how much schedule risk the big buyers are willing to tolerate right now: none.
Read Crusoe's own post alongside TechCrunch's write-up; the price and the reasons for the switch come from reporting, not from either company.
TechCrunch: https://techcrunch.com/2026/11/06/crusoe-sells-troubled-wyoming-data-center-project-to-google/
Crusoe: https://www.crusoe.ai/resources/newsroom/crusoe-wyoming-data-center-campus-update
← Back to all articles
The backstory is the interesting part. Crusoe announced the site last year with Blackstone-backed Tallgrass, planning to scale it to 10 gigawatts. It tried and failed to land Google as the customer. Google reportedly raised concerns about cost and construction timeline under Crusoe's management, Bloomberg first reported the handover in June, and now Google is simply building it itself. The two still work together: Crusoe is developing another campus for Google near Amarillo, Texas.
None of this stopped Crusoe's momentum. It raised $3.9 billion in September, builds the Abilene, Texas campus for Oracle and OpenAI, is building a 900-megawatt campus next door for Microsoft due mid-2027, counts Meta as a customer, and per Bloomberg signed a roughly $13 billion, five-year deal with Jane Street.
Why agent people should care: inference for agents runs on exactly this kind of capacity, and the bottleneck is increasingly delivery, not demand. Power, equipment and local opposition decide when compute comes online. A hyperscaler that thinks a developer is too slow will just take the keys, which tells you how much schedule risk the big buyers are willing to tolerate right now: none.
Read Crusoe's own post alongside TechCrunch's write-up; the price and the reasons for the switch come from reporting, not from either company.
TechCrunch: https://techcrunch.com/2026/11/06/crusoe-sells-troubled-wyoming-data-center-project-to-google/
Crusoe: https://www.crusoe.ai/resources/newsroom/crusoe-wyoming-data-center-campus-update
Comments