Increasing worry for datacenter managers regarding capacity due to AI expansion
In the ever-evolving world of datacenter operations, the Uptime Institute's 15th Annual Global Data Center Survey (2025) has revealed a complex landscape marked by significant challenges and uncertainties.
The report highlights that forecasting future capacity requirements is the second biggest concern for datacenter operators, surpassed only by cost concerns. This is due to the increasing adoption of AI workloads, which are resource-intensive and consume heavy electrical power.
Operational expenses are also on the rise, primarily due to escalating energy prices and market volatility. Efficiency requirements and regulatory pressures add further complexity to the planning process.
Staffing shortages and supply chain disruptions continue to impact datacenter operations and expansion plans, complicating capacity scaling and modernization efforts. In fact, 46% of respondents are having difficulty finding qualified candidates for vacant roles, and senior roles in operations management are particularly hard to fill.
Despite these challenges, the report also indicates a relentless growth in public cloud infrastructure. However, 45% of IT workloads remain on-premises, with 16% in colocation, showing that many organizations still manage significant in-house capacity that must be planned for accordingly. Public cloud accounts for about 11%, with hosted private cloud and SaaS making up additional portions.
Simultaneously, AI technologies are transforming datacenters by enabling smarter resource optimization, predictive maintenance, and self-adaptive management. AI-driven automation can dynamically allocate resources and improve energy efficiency, helping address capacity and cost issues.
Operators are cautious in their adoption of AI, mostly interested in tools that can increase facility efficiency or help to reduce human error. Encouragingly, 73% of respondents trust an adequately trained AI model to run automated analytics on sensor data.
Despite the challenges, the industry is making slow but steady progress. The average server rack power density, excluding outliers, is 7.5 kW, up from 6.8 kW in 2024. Moreover, only 3% of operators experienced the most severe kind of incident, and 50% of datacenters had at least one outage in the past three years, a decrease from last year's 53%.
However, the industry faces a management shortage, with organizations failing to pass on knowledge to more junior staff as senior workers leave the workplace. This knowledge gap could potentially hinder future growth and innovation.
In conclusion, datacenter capacity planning in 2025 is a complex, high-stakes challenge, shaped by the competing priorities of supporting cutting-edge AI workloads, controlling escalating costs, overcoming staffing and supply chain difficulties, and navigating the industry’s gradual but persistent transition toward hybrid cloud models. Operators must adopt new AI-powered tools and planning approaches to achieve resilience, efficiency, and scalability in this evolving environment.
- Datacenter operators are concerned about anticipating future capacity requirements, as the adoption of AI workloads, known for their resource-intensive nature and high electrical power consumption, increases.
- The report indicates a rise in operational expenses due to escalating energy prices, market volatility, efficiency requirements, and regulatory pressures.
- Staffing shortages and supply chain disruptions continue to pose challenges for datacenter operators, making it difficult for them to scale and modernize their operations.
- The Uptime Institute's survey shows that while there is a growth in public cloud infrastructure, many organizations still manage significant in-house capacity, highlighting the need for software and hardware solutions that can adapt to data-and-cloud-computing and technology advancements, such as AI-driven automation for resource optimization and predictive maintenance.
