Google buys Crusoe's data center in Wyoming for $17 billion

Crusoe, a Denver-based data center developer valued at $30.9 billion, has confirmed the sale of an unfinished project in southeastern Wyoming to Google. The Mountain View giant, which ironically had courted Crusoe as a customer for this same facility without success, is now taking over the development with plans to invest $17 billion in the campus.

The exact amount paid for the acquisition has not been disclosed. However, the context of the negotiation speaks volumes about the current state of the infrastructure market for artificial intelligence: even billion-dollar companies are hitting the physical and logistical limits of building at the speed that computing demand requires.

What went wrong with Crusoe's project in Wyoming

Crusoe had announced the project last year in partnership with Tallgrass, an infrastructure company backed by Blackstone. The original plan was ambitious: to scale the facility to 10 gigawatts of energy capacity. To put that number in perspective, most large data centers currently in operation work with much lower capacities.

According to reports, Google raised concerns about the cost and construction timeline under Crusoe's management. Securing equipment, ensuring power supply, and dealing with increasing opposition from local communities made delivering on time and within budget a task that the startup could not accomplish on its own.

This is a recurring pattern in the sector. Independent data center developers face bottlenecks ranging from the supply chain of transformers and generators to regulatory disputes with municipalities and utility companies. Crusoe's case in Wyoming is emblematic, but far from isolated.

Why Google decided to take over instead of abandoning

Google's decision to buy the project instead of simply seeking alternatives reveals a central dynamic in the AI infrastructure market: land with access to large-scale energy is a scarce asset. Finding a new site, negotiating connections to the power grid, and obtaining environmental permits can take years. Taking over an already underway project, even if problematic, shortens that path.

The $17 billion that Google plans to invest in the Wyoming campus adds to the company's global offensive in infrastructure. The race for computing capacity to train and run AI models has turned data centers into the most sought-after asset in the technology sector.

Despite the setback in Wyoming, the relationship between Google and Crusoe has not ended. The startup continues to develop another data center for Google near Amarillo, Texas. This suggests that the issue was not a lack of trust in the company as a whole, but rather its execution capacity on that specific project.

Crusoe: from Bitcoin mining to major AI contracts

Crusoe's trajectory is, in itself, a portrait of the transformations in the technology sector in recent years. Founded by Chase Lochmiller and Cully Cavness, the company started as a cryptocurrency mining business powered by natural gas that would otherwise be flared and wasted. The model of harnessing waste energy to mine Bitcoin was ingenious, but the real leap came with the migration to AI infrastructure.

Today, Crusoe operates on a completely different scale. Its campus in Abilene, Texas, is being built in phases for Oracle and OpenAI. A second 900-megawatt campus in the same region serves Microsoft, with the first building expected to be operational by mid-2027. Meta is also a client. And, according to market information, trading firm Jane Street signed a $13 billion, five-year cloud services contract with Crusoe.

In September, the company raised a $3.9 billion round, even with the Wyoming issue already public. Among the new board members is JB Straubel, co-founder of Tesla and founder of Redwood Materials, who had invested in Crusoe in 2021. Crusoe has even become the first customer of Straubel's energy storage business.

What this means for the data center market

The Crusoe-Google episode illustrates a structural tension. The demand for computing for AI is growing at an exponential rate, but the construction of physical infrastructure remains constrained by real-world limitations: construction time, energy availability, equipment manufacturing capacity, and community acceptance.

Large companies like Google, Microsoft, and Amazon have the balance sheets to absorb delays and budget overruns. Startups, even those valued in the tens of billions, operate with much smaller margins for error. Crusoe's business model, which generates revenue by renting space in data centers, leasing its own GPUs, and selling computing power for AI model inference, is solid in theory. But executing projects at the scale of 10 gigawatts requires a level of operational maturity that few independent players possess.

For the investor tracking the technology sector, the message is clear. AI infrastructure is the new oil in terms of demand, but the costs and risks of construction are real and often underestimated. Companies that can deliver projects on time and within budget will have a lasting competitive advantage. Those that cannot end up selling assets to those who can.

Crusoe's Wyoming is now Google's Wyoming. And the $17 billion that follows shows that, in this race, capital is not the problem. Execution is.