AI and Cloud – Where is the Value

Observations

All Cloud Service Providers (CSPs) are looking to meet the AI challenges and in doing so are seeing increases in supply costs associated with upscaling their compute and storage infrastructure, including new AI-optimised chipsets, and additional environmental costs associated with powering and cooling of data centres.

This change has caused CSPs to quietly change the dynamics of their cost models and to reassess what can make them the supplier of choice for your business.

The introduction of AI also should make all customers more aware of data sovereignty – which in turn CSPs are aware of and where this has not been a part of their offering, is now becoming so.

Both Azure and AWS have offered data sovereignty for some time, now Google is looking to gain market share and has also developed its AI offering with data, training, and AI processing sovereignty in the mix.

The issues of Cloud being a pseudo lock-in arrangement with suppliers has also started to change. The recent case study on Orica’s¹ move of its enterprise resource planning (ERP) Software-as-a-Service (SaaS), from one Cloud to another demonstrates that a change in CSP, even when you see the SaaS provider as the primary contracted provider, can deliver benefits in both performance and running costs. Even where your primary platform is M365, both the major competitors AWS and Google offer the ability to house the M365 platform on their Cloud service.

So what should your organisation do to keep the use of Cloud under control from a data sovereignty, AI, and consumption cost point of view?

IBRS recommends six areas to regularly review to ensure you have the best value proposition from your organisation’s use of Cloud relative to costs and productivity in the AI world. These are:

• Do your SaaS suppliers offer flexibility in the use of Cloud services? Flexibility in the use of Cloud suppliers can lead to significant improvements in performance and enable better cost management without requiring substantial changes to the way business areas consume these services.
• Can you effectively calculate and accurately predict the full cost of your Cloud enterprise? The ability to calculate your costs for your Cloud enterprise will provide transparency into the costs associated with networking, compute, storage, and recovery options, as well as whether the use of containerisation may benefit your business’s legacy applications.
• Do you have effective monitoring of your Cloud consumption costs? Effective monitoring of Cloud consumption costs is essential. If you are not running a Cloud FinOps model² you should consider its introduction. It will enable transparency for the business on the value of Cloud computing for the organisation and allow for data-driven decisions.
• Is the use of Cloud for data storage efficient? Is your Cloud storage efficient and usable? Too often, many organisations have transferred their on-premise thinking of how and what to store into a Cloud environment without reviewing how Cloud can streamline the solution. Consider the purpose of the data store: online access, backups (including what needs to be backed up and retention policies), and immutable backups for ransomware protection.
• Is data sovereignty important to your organisation? Digital data sovereignty extends beyond the location of data to encompass the placement of all compute services, including AI, as well as the location of support teams that can access your Cloud services and data. It also includes storage and control over encryption keys, among other factors. How can you gain a cost advantage from other Cloud suppliers relative to data sovereignty? How does your internal use of AI tools impact the data holdings? Do you need to isolate data?
• Where does embedded AI offer an advantage? AI can be a targeted and specific tool for business, but embedded AI functionality in your Cloud platform may also offer a productivity gain. Is your current CSP offering competitive?

IBRS recommends that CIOs review the cost-effectiveness of their Cloud usage using a FinOps approach. Not just in dollar terms, but also in terms of productivity, and data security in the use of AI and in doing so, the need to protect data. While the tier one Cloud suppliers are unlikely to change in the near term, their adoption of AI will lead them to modify their pricing models and seek to attract customers with improved functionality and the adoption of data sovereignty. An annual review will see your organisation stay ahead of the curve.

IBRS recommends that, before deciding to change suppliers, if that is the outcome of your review, it is essential to conduct a proof-of-value exercise to demonstrate that the productivity gains and cost reductions can be achieved, taking into account the cost of change in terms of both internal and external effort.

Next Steps

• If one is not already in place, begin working towards establishing a Cloud FinOps model.
• Review your current exposure to Cloud Services, both through direct CSP agreements and via SaaS providers.
• Conduct proof-of-value exercises to determine whether the potential change in Cloud suppliers delivers benefits.
• Develop your business case for change – demonstrate that the change will deliver benefit.
• Conduct regular (at least annual) reviews of Cloud (CSPs and SaaS) to ensure your organisation is getting value.

Footnotes

1. ‘VENDORiQ: Yes! An ERP Cloud Migration Can Unlock a Data Advantage!’, IBRS, 2025.
2. ‘FinOps Principles’, FinOps Foundation, 2025.