Public cloud is not the final destination. So what is?

The public cloud is not the final destination for every application or workload, nor will it be in the foreseeable future.

Only 13% of organisations say they have 100% public cloud IT environments, according to 451 Research’s Voice of the Enterprise: Cloud, Hosting & Managed Services and Cloud Native, FinOps 2024 survey. Enterprise workloads have different resource, performance, security, compliance and data sovereignty requirements — all good reasons for keeping some applications on-premises.

The wave of recent AI-related activity drives this point home. Organisations often leverage
public cloud for open-source and public AI models, pipeline tools, and some training and
inferencing. However, cost considerations, data governance, privacy and security concerns, and
the emergence of distributed edge application requirements mean that there is often value to be found in private cloud/on-premises deployments.

Where does a hybrid cloud operating model come in?

A hybrid cloud operating model is critical for harnessing modern workloads, including AI, that span multiple infrastructure environments. IT heterogeneity is not a new phenomenon, and enterprise IT estates can take many forms.

Complexity is the result as successive generations of typically siloed technology and infrastructure accumulate, reflecting a range of previous decisions and priorities. Organisations with these “hybrid-by-accident” IT environments often face steep challenges, particularly when public and private cloud enter the mix. This frequently results in “swivel chair” tooling and operations management, limited visibility across the IT estate and multiple vectors of inefficiency.