Crusoe sells troubled Wyoming data center project to Google

Introduction

Crusade, the Denver‑based data‑center developer, has agreed to sell its stalled Wyoming project to Google. The transaction, first reported by Bloomberg, highlights the growing pains of independent firms trying to deliver ultra‑large AI‑infrastructure on tight timelines and budgets. While Google steps in to finish the campus, Crusoe continues to work on other sites, including a Texas facility for the search giant.

Why the Wyoming project faltered

Crusoe announced the Wyoming campus last year with Tallgrass, a Blackstone‑backed infrastructure partner, aiming for a 10‑gigawatt power capacity. The ambition quickly ran into three common obstacles:

1. Cost overruns – Building a data center of this scale requires massive upfront capital for power, cooling and networking. Crusade’s estimates proved optimistic, prompting Google to question the financial viability under Crusoe’s management.
2. Construction timeline – Securing equipment, especially high‑density GPU racks, has become a bottleneck across the industry. Delays in delivery pushed the project past its target opening date.
3. Local resistance – Community concerns about land use, water consumption and environmental impact can stall permitting and add regulatory friction.

These factors combined to make the Wyoming site a liability for Crusoe, prompting the decision to hand the unfinished assets to a buyer with deeper pockets and more experience navigating large‑scale builds.

What Google gains from the acquisition

Google has not disclosed the purchase price, but it announced a $17 billion investment to complete the campus. The acquisition offers several strategic benefits:

• Control over a key geographic node – Wyoming’s low‑cost power and ample land fit Google’s long‑term plan to diversify its global footprint.
• Accelerated AI capacity – By taking over an existing, partially built site, Google can bring new compute resources online faster than starting from scratch.
• Supply‑chain leverage – Google’s scale gives it preferential access to GPU manufacturers and cooling technology, reducing the risk of future component shortages.

Crusoe’s broader outlook

The Wyoming sale does not signal an exit from Google entirely. Crusoe remains the developer of a new Texas data‑center for the search giant, and it continues to raise capital—most recently a $3.9 billion round in September. The company’s business model blends traditional colocation (leasing space to customers who bring their own GPUs) with its own GPU‑rental and AI‑inference services. Its flagship projects include:

• Abilene, Texas – A multi‑phase campus serving Oracle and OpenAI.
• Adjacent Microsoft campus – A 900‑megawatt facility slated for mid‑2027 activation.
• Meta and Jane Street – High‑profile customers that underscore Crusoe’s relevance in the AI‑compute market.

Leadership changes, such as the addition of JB Straubel to the board, signal a continued focus on energy‑efficient operations, leveraging Straubel’s expertise in battery storage and renewable integration.

Practical takeaways

• Independent developers must align cost, schedule and community engagement early to avoid costly hand‑overs.
• Large cloud providers like Google can absorb unfinished projects when they align with broader capacity goals, turning a risk into a strategic asset.
• Diversified revenue streams—colocation, GPU leasing, and inference services—help developers weather setbacks in any single project.

Conclusion

Google’s takeover of Crusoe’s Wyoming data‑center underscores the growing complexity of building AI‑grade infrastructure at scale. While the move resolves a stalled venture for Crusoe, it also reinforces the advantage that deep‑pocketed cloud operators have in navigating cost, supply‑chain and regulatory challenges. For independent developers, the lesson is clear: robust planning, realistic budgeting, and proactive community outreach are essential to stay afloat in a market where the stakes—and the power demands—are higher than ever.