Rising Costs and Supply Constraints in Cloud Computing Power

Nebius, listed on the NASDAQ as NBIS, is positioning itself to navigate the ongoing shortage of AI graphics processing units (GPUs). This shortage has created a volatile market for computing power used in cloud services, characterized by reduced supply and increasing prices.

Market dynamics are shifting as companies seek alternatives to traditional cloud providers. This environment has led to the establishment of new contracts with non-cloud solution providers to secure necessary hardware and processing capabilities.

The Infrastructure Investment Race

The push to scale data centers to meet AI demands has evolved into what McKinsey describes as a $7 trillion race. To manage this scale, organizations are prioritizing investments in computing technologies that are both cost- and energy-efficient.

The goal of these investments is to optimize performance while simultaneously managing infrastructure costs and power consumption, as the sheer volume of compute required for AI continues to grow.

Power Constraints and Grid Stability

While hardware availability is a primary hurdle, the physical infrastructure required to power that hardware is becoming a critical bottleneck. Data center operators are increasingly concerned about power constraints.

According to reporting from Business Insider, rising costs were the chief concern for data center operators in 2024, but concerns regarding the power supply are now quickly catching up to financial costs.

The National Academies of Sciences, Engineering and Medicine are examining how the electrical grid can keep up as electricity demand grows and associated risks increase. This coincides with research from the Pew Research Center regarding the energy use of US data centers amid the artificial intelligence boom.

Financial Resilience and Risk Management

The rising cost of cloud services is forcing companies to rethink their financial strategies. Akamai has noted that while reducing cloud spend is a priority, the broader needs of financial resilience and risk management transcend simple cost considerations.

The necessity for massive processing power to support cloud operations means that companies must balance the immediate cost of compute against the long-term risk of inadequate capacity.

As the industry navigates the AI GPU shortage, the intersection of hardware procurement, energy availability, and financial risk management will determine which providers can successfully scale their operations.