Oracle’s data center ambitions are reportedly running into major cost surprises, underscoring how expensive the race to build AI infrastructure has become. The latest example centers on a proposed $165 billion project in New Mexico, where the company is said to be trying to rescue plans that have run into trouble.
The report points to a broader challenge facing companies building large-scale AI campuses: the final price tag can grow far beyond what was first expected. As demand for computing power rises, cloud providers are under pressure to expand quickly, but massive facilities require enormous upfront spending and can become more complex and costly over time.
Oracle had earlier this year been preparing to commit billions of dollars, according to the report, but the New Mexico development now highlights how unpredictable these projects can be. Even for major technology companies, building out advanced data center capacity appears to involve financial risks that are difficult to fully anticipate at the outset.
The situation reflects a wider trend across the industry as companies chase the infrastructure needed for AI services. Large data center investments may promise long-term strategic value, but Oracle’s reported experience suggests that scale alone does not protect projects from budget shocks and execution problems.
