As a revolution around generative AI (GenAI) looms, investment in data centres is expected to rise dramatically this year as cloud providers dedicate significant financial resources to expanding facilities to support cloud services and AI-driven applications, says GlobalData.
The research group’s latest report – 2025 Enterprise Predictions: Data Centre – reveals that in 2025, colocation providers like Equinix, Vantage, NTT, Digital Realty, and Vertiv will continue to grow driven by increasing demand for AI services by customers in numerous industries. These companies are expected to expand their real estate using economies of scale to scout increasingly harder-to-find data centre locations.
Beatriz Valle, senior analyst for Enterprise Technology and Services at GlobalData, says: “Enterprises will continue to resort to co-location providers to deploy their AI infrastructure rather than build their own facilities. This eases the complexity associated with deploying and operating AI-driven workloads at scale and outsource any liabilities associated with the upkeep of the data centres.
“Co-location facilities provide enterprises with not just power and cooling – and a safe location for their equipment – they also connect those systems to a range of telecommunications and network service providers at a fraction of the cost,” Valle adds. “Enterprises are looking at co-location data centres for not only AI, but also for bringing workloads that are no longer economical out of the public cloud and into their own private one, hosted at the ‘colo’.”
GlobalData predicts on-premises workload deployments will endure, and a massive cloud exodus remain elusive. Although most organisations continue to migrate workloads to the cloud to benefit from digital transformation and application modernisation initiatives, it is now becoming apparent that some workloads are not designed for cloud computing deployments.
“Put simply: some critical workloads in specific industries are not designed for the cloud,” says Valle. “Industries managing particularly sensitive data such as finance, healthcare, life sciences, and some government sectors face up to so much complexity it is not worth the trouble. In addition, some geographies have stricter regulation and compliance, creating the need for data sovereignty.”
On-premises will continue to offer certain advantages such as greater security, control, integration with existing IT operating systems, and lower latency for applications requiring quick data access. In addition, financial considerations such as unpredictable fees will mean that on-premises will continue to prove resilient going forward.
“Moreover, there will be anecdotal, but consistent evidence that the advantages mentioned above will move some organizations – usually large enterprises with specific applications – to bring their public cloud workloads back to a private cloud environment,” Valle says.