10 trends for datacenters in 2025

We never tire of repeating it. As datacenters develop rapidly and carbon neutrality is now a global mission, the sector is undergoing unprecedented changes. In addition to high reliability and efficient energy use, the requirements for sustainable datacenters today include greater use of renewable energy, smarter management, faster deployment, more elastic capacity expansion and more advanced overall sustainability.

Huawei recently published a paper revealing the top 10 trends in datacenter installations based on its customers’ perceptions. They are:

• Low-carbon green energy will be more widely used – Efficient energy consumption will reduce PUE; the rate of energy recovery will increase; more innovative energy conservation technologies will be applied, such as waste heat recovery.

• Sustainability – High energy efficiency is a prerequisite for sustainable development. WUE will become more important. Datacenters will have less impact on the surrounding environment.

• Fast implementation – Time to market (TTM) will drop from 12 to 6 months or even be shorter.

• High-density datacenters – which will use integrated, high-density power supply systems. High density will drive the evolution of liquid cooling technology.

• Elasticity – One generation of facilities will accommodate two or three generations of IT equipment. Deployment will be flexible to cater for different power densities.

• Distributed cooling – The reliability of the cooling system in a datacenter is often overlooked, despite the fact that it is responsible for 13% of datacenter accidents or outages – the third biggest source of failures after power and network failures, according to recent research data.

• Intelligent O&M – Datacenter facilities will be even more digitised, and all links will be visualised, manageable and controllable. AI is fast becoming an essential tool for datacenter operations and management.

• Security and reliability – Predictive maintenance at system, component and device level is necessary to continuously improve hardware reliability and system resilience.

Technological advances, sustainability efforts and the growing importance of data security are shaping the future of datacenters. As these trends continue to evolve, companies must adapt their strategies to ensure infrastructure stability, support business performance and remain competitive.

According to Forrester’s recently released Budget Planning Guide 2025: Technology Executives, 91 per cent of the technology decision-makers participating in the study predict an increase in IT budgets for the next fiscal year. Computing and networking hardware accounts for 13.3 per cent of total IT spending, according to ISG’s global benchmarking data. As architectures change, it is essential to maintain a stable base for business performance with a variety of enablement options. For example, according to Forrester’s Q2 2024 Tech Pulse Survey, 31% of technology decision-makers identify inadequate technology infrastructure as the biggest obstacle to improving their IT organisation’s speed to market.

Public cloud investments will grow in 2025 to support AI infrastructure. Many large companies have been using AI for specific use cases for several years; often, they have utilised fully managed services for machine learning as they have expanded their overall cloud infrastructure. Today, Generative AI accounts for a larger share of overall cloud spending as users move from experimentation to production at scale, even as the shift to cloud-native services such as Kubernetes and serverless continues.

The growing demand for energy to run AI and GenAI workloads is contributing to rising CO2 emissions. Cloud platforms and services are already responsible for more global greenhouse gas emissions than the airline industry. Regulations such as the EU’s Corporate Sustainability Reporting Directive and the SEC’s Climate Disclosure Rule require organisations to report their carbon footprint.

Technology leaders should invest in AI tools that enable intelligent allocation of workloads in datacenters based on needs and even energy sources, such as Google Cloud region picker or Azure carbon optimisation and Microsoft Azure emissions insights in Microsoft Fabric. They can also implement capacity optimisation technologies, such as those offered by CAST AI and the open source Cloud Carbon Footprint, to reduce the overall carbon footprint of their cloud operations.