Crusoe contends to have contracted 4.9 GW of contracted pipeline across data center projects and Crusoe Cloud, with a “total development pipeline” that the company says exceeds 40 GW (including contracted projects, sites under active tenant negotiation, and projects in advanced development).

According to a recent interview RCRTech conducted with Crusoe, a significant portion is being built or being actively deployed, and the most recent aerial view of its Abilene, TX campus reveals a lot of data centers are being built. But, it begs the question of “How” at this scale can that much capacity be moved from contracted GW to actual energized, operable GWs, given the shortages of substations, transformers, and switchgear, and the challenges of cooling and providing consistent compute capacity under real-world conditions?

Immediate viability faces real-world bottlenecks, like the supply chain issues I raise above, and some hesitancy by its own customers and those allocating capital to massive infrastructure projects.

For example, the 1.2 GW Stargate site near Amarillo/Claude will be built in phases, with the first two buildings operational and six more under construction. The part of the Abilene campus that was originally purpose built for Oracle, with plans for 1.2GWs spanning eight buildings and nearly 4 million square feet, is now being expanded by Microsoft, since Oracle and OpenAI scaled back their plans.

Wyoming’s Project Jade is on an indefinite pause, with Crusoe recently pulling out.

To go beyond the demand into actual execution of 4.9GWs, Crusoe will count on its 2.1 GW Abilene, Texas Hub, and three or so other undisclosed hyper-scaler projects across Texas and Missouri.

With the volatility of customer demand, community pushback, and capital allocation, as well as grid interconnection issues, it won’t be easy to turn that 4.9GW into energized, commissionable, operable GWs. However, alternative power setups (i.e., behind-the-meter gas and fuel cells), and a powerful combination of multi-year hyperscale builds and rapid factory-built deployments, like modular “Crusoe Spark” units for faster inference, will perhaps make it viable.

RCRTech has reached out to Crusoe for further comment on this story, and will report as it unfolds.

Below, check out highlights from our on-demand webinar, “Scaling optical networks for the AI and Hyperscale era” with Nokia, Ciena, Dell ‘Oro, and AFL; and, take a look at today’s data center news in “What you need to know.”

Webinar: scaling optical networks: Spending patterns in the optical market are increasingly shaped by hyperscalers rather than telecom operators – says Ciena, Nokia, and Dell’Oro in a recent RCRTech webinar about AI as a driver of optical.

Alibaba expansion in SE Asia: Alibaba Cloud has expanded its Malaysia footprint with 5 data centers–two of which are in the Johor region. There are plans to introduce agentic AI services for enterprise customers later this year.

350MW campus in Norway: atNorth spoke to RCRTech about NOR01, its planned 350MW campus in Norway, which will be designed for AI and HPC workloads, and will tap the abundant renewable energy and power networks available in the region.

769-acre, 400MW Fleet DC: Fleet Data Centers broke ground on its 2nd data center campus at Tahoe Reno Industrial Center, with plans for two 200MW data centers – one at South Valley and the other at Peru Ridge.

NY semiconductor corridor: 350 industry leaders, researchers, manufacturers, startups, investors and government partners gathered for the NY SMART I-Corridor Semiconductor Summit, highlighting the regions embrace of the semi ecosystem.

CT comprehensive AI legislation: CT Gov. Ned Lamont signed Senate Bill (SB) 5 into law, establishing a framework for AI tech, incl. subscription-based AI services, frontier models, AI companions for minors, AI in hiring, and watermarking of content.

Nebius expands in UK: Nebius committed approximately £1.7 billion to capacity buildout at 4 U.K. sites, with 3 deployments of NVIDIA full-stack, end-to-end AI factory platform technology as it expands its commercial and AI R&D hub in London.

Applied Digital secures up to $550M: Applied Digital closed a $350 million revolving credit facility and possible accordion option of up to $200 million, arranged by Goldman Sachs, for the pre and post-lease development of its data centers.