Crusoe Secures $11.6 Billion to Expand Texas AI Data Center, Supporting OpenAI Infrastructure

AI infrastructure startup Crusoe has raised an additional $11.6 billion to significantly expand its upcoming data center in Abilene, Texas, marking one of the largest funding rounds in the emerging “neocloud” space. The new capital brings the total raised for the project to $15 billion and will allow Crusoe to expand the facility from two to eight buildings, the company confirmed on Wednesday.

Founded in 2018 as a crypto-focused firm, Crusoe has since pivoted to become a specialized cloud provider for AI workloads, part of a new wave of “neoclouds” that offer tailored infrastructure beyond the traditional giants like AWS, Azure, and Google Cloud.

Crusoe has been contracted by Oracle to construct the first data center for Stargate — a major AI infrastructure initiative backed by OpenAI, SoftBank, and Oracle, with a planned $500 billion investment in global AI infrastructure. According to The Wall Street Journal, the Abilene facility is set to become OpenAI’s largest data center.

“Our customer is Oracle. OpenAI is Oracle’s customer,” Crusoe clarified in a statement, emphasizing its indirect yet vital role in supporting the ChatGPT creator’s infrastructure needs.

The project is seen as part of OpenAI’s long-term goal to reduce reliance on Microsoft, its current primary cloud provider.

Key Details:

  • Location: Abilene, Texas

  • Total Buildings: 8 (up from 2)

  • AI Chips: Each building will house up to 50,000 Nvidia Blackwell systems

  • Sponsors: Crusoe, Blue Owl’s Real Assets platform, and Primary Digital Infrastructure

The facility will support intensive generative AI workloads, crucial for OpenAI’s future model development and deployment.

The explosive growth in demand for AI compute capacity has fueled an investment boom in data centers powered by specialized chips like Nvidia’s Blackwell series — a market Crusoe is aggressively entering.

Neither OpenAI nor Nvidia responded to requests for comment at the time of publication.