AI cloud infrastructure provider Nebius Group reported a sharp increase in capital expenditures in the latest quarter as it accelerates spending on graphics processing units and data center expansion to meet rising artificial intelligence demand. The company’s aggressive investment strategy highlights the ongoing race among cloud providers to secure computing capacity in a supply constrained environment.
Nebius said capital expenditures climbed to approximately 2.1 billion dollars in the December quarter, compared with 416 million dollars in the same period a year earlier. The surge was primarily driven by purchases of advanced Nvidia processors and expanded data center infrastructure across multiple regions.
The Amsterdam based company, which counts major technology firms such as Microsoft and Meta among its customers, plans to expand its footprint with nine new data center sites in the United States, France, Israel and the United Kingdom. The expansion reflects sustained enterprise demand for AI training and inference workloads.
Revenue for the fourth quarter rose more than six fold to 227.7 million dollars, underscoring the rapid scaling of the business. However, the figure fell short of analyst expectations of roughly 246 million dollars. Despite the revenue growth, Nebius reported a widening net loss of 249.6 million dollars, compared with 133.2 million dollars a year earlier, as heavy investment spending weighed on profitability.
Nebius operates as a neocloud provider, offering hardware and AI optimized cloud capacity to technology companies. Its core business centers on providing access to Nvidia GPUs and specialized AI infrastructure. Alongside rivals such as CoreWeave, the company has benefited from continued corporate investment in artificial intelligence applications ranging from generative AI models to large scale data analytics.
Management said demand from enterprise and AI native customers continues to exceed available supply, allowing the company to secure contracts for future capacity well in advance. Nebius has already contracted more than 2 gigawatts of power and expects to exceed 3 gigawatts by the end of the year, surpassing earlier guidance.
The rapid buildup of power capacity signals the scale of infrastructure required to support AI workloads, which are significantly more energy intensive than traditional cloud computing services. Securing long term power agreements has become a strategic priority for AI cloud providers as they compete for prime data center locations and reliable electricity access.
Looking ahead, Nebius projects an annualized revenue run rate of between 7 billion and 9 billion dollars by the end of 2026, up from 1.25 billion dollars at the close of 2025. Achieving that target will depend on continued enterprise spending on AI and the company’s ability to deploy new capacity efficiently.
Shares of Nebius, which surged more than 200 percent last year, were volatile in trading following the results. Investors are weighing strong top line growth and long term demand prospects against mounting capital expenditures and near term losses in a highly competitive AI cloud market.
