The artificial intelligence boom is creating a new class of winners.
While investors remain focused on companies building AI models and applications, a massive race is unfolding behind the scenes to secure the computing power needed to run them. The latest beneficiary is Argentum AI, which has secured $7.8 billion in long-term infrastructure contracts tied to one of Europe's largest AI GPU deployments.
Backed by Super Micro Computer, Argentum plans to deploy approximately 47,000 Nvidia GB300 GPUs across Poland, placing the project among the largest AI infrastructure rollouts currently underway in Europe.
The announcement comes as global spending on AI infrastructure continues to accelerate, driven by surging demand for AI training, inference, and enterprise AI applications.
A Multi-Billion Dollar Bet on AI Compute
The largest agreement is a $2.5 billion partnership with cloud gaming provider Boosteroid and real estate developer DL Invest Group.
The deal centers around a 300-megawatt AI data center in Bielsko-Biała, Poland, designed to support tens of thousands of Nvidia GB300 and Blackwell GPUs. The facility will serve AI model training, inference workloads, and enterprise cloud computing services.
Argentum previously announced a separate $1.5 billion agreement for dedicated GPU cloud infrastructure equivalent to approximately 10,500 Nvidia GB300 GPUs.
Combined with additional contracts, the company says its total commitments now stand at $7.8 billion.
Loading tweet...
View Tweet
The Real AI Race Is About Infrastructure
The significance of the announcement extends far beyond one company.
AI demand is creating an unprecedented scramble for GPUs, data centers, networking equipment, and power infrastructure. While firms like OpenAI, Anthropic, xAI, and Google compete to build more powerful models, they all rely on the same underlying resources: compute.