Business technology leaders are expecting their cloud spending to grow in 2024 due to the growth of AI. Cloud economics is an intelligent management strategy that can better manage these costs as generative AI inevitably pushes spending higher.

AI applications are only as good as the data they use. The more high-quality data that algorithms have access to, the more accurate and reliable their results will be.

But just how much data do machine learning and large language models require? As AI use cases become more complex and resource-intensive, the expenses associated with supporting them—and particularly the cloud infrastructure for AI workloads are increasing.

Unfortunately, there’s no hard-and-fast rule when it comes to forecasting what's needed to run AI models. Workload requirements and computing power vary depending on the size and complexity of the project. Natural language processing, for example, is a complex task that typically requires a lot of data for training. It also happens to be one of the more popular use cases, with companies using it for everything from email generation and document summaries to customer service. These tasks can quickly become even more complex if a business wants highly contextualized responses, such as employee self-service for a global company or help handling sensitive information in a highly regulated industry such as healthcare or finance. These generative AI models can run "10 to 100 times bigger than older AI models"; meanwhile, use cases and user bases are growing.

AI’s data demand is having a major impact on cloud costs at a time when many companies are scrutinizing cloud budgets. Early lift-and-shift approaches to cloud adoption have led to bloated cloud budgets, with many companies paying for cloud services that they don’t need. Now that the dust surrounding these migrations is settling, companies are taking stock of their cloud consumption and taking steps to reduce overspending. However, with the explosion of AI, these cloud infrastructures must now be redesigned to handle large-scale AI use cases.

This might seem like terrible timing: A promising, data-intensive technology goes mainstream just as companies are tightening their data and cloud infrastructure belts. But growing interest in cloud economics is showing companies how to maximize the value of new technology investments rather than cutting them out altogether.

Applying Cloud Economics To Manage AI Costs

Cloud economics is a strategic approach to managing cloud spending that goes beyond simply cutting costs or increasing efficiency. It involves conducting a thorough cost-benefit analysis to ensure that cloud investments align with business priorities and will maximize overall business value. This same approach can be applied to managing AI costs.

To effectively manage AI costs using cloud economics, businesses need to consider the potential impact and business value of the new technology. It’s crucial to align AI investments with broader business objectives and understand the accommodations required for successful implementation. Instead of investing in AI use cases based solely on perceived value, teams should work to understand what “valuable” means for the business and set clear expectations for how to achieve that value.

All business teams that intend to leverage AI models should collaborate within a shared framework and process for approving AI-driven projects. When teams work independently to make decisions about cloud and AI investments, it becomes challenging for the business to assess the costs and benefits of those decisions. Establishing a centralized cloud economics team or a group within a cloud center of excellence (COE) can help coordinate these efforts, establish best practices and guide teams on spending, experimentation and investment.

Additionally, innovation COEs can play a key role in helping businesses understand the requirements and value of new technologies, including AI. With AI advancing rapidly, businesses must grasp the potential impact and opportunities these advancements bring. By leveraging cloud economics to manage AI costs, businesses can help ensure that their investments align with strategic objectives and deliver maximum value.

Optimizing AI Workloads With Cloud Economics

With a solid cloud economics program in place, businesses can delve into usage data to truly optimize performance across AI and the cloud. Initially, this involves reducing wasted spending by shutting down underutilized resources. Cloud economics can help businesses take advantage of variable cost models to support AI investments. AI data requirements tend to be high, but they’re also variable. Businesses stand to save a lot of money by accurately forecasting cloud requirements and adjusting as needed. This requires a strong data analytics program capable of generating timely insights. These adjustments can free up resources, which can then be reallocated to fund more critical (AI) investments.

Teams can further utilize performance data to determine the optimal cloud environments for their enterprise, incorporate intelligent workload management to automate resource scheduling and adjust cloud services based on AI demand.

Using cloud economics to manage AI costs not only helps ensure effective cost management but it also enables businesses to optimize the AI workload performance. By leveraging usage data and variable cost models, businesses can make informed decisions to maximize the value of AI investments while aligning them with broader business objectives. This strategic approach will help reduce overspending and allow businesses to reallocate resources effectively, realizing the full potential of their cloud and AI investments.

This article was originally published on

About the Author


Jo Debecker

Jo Debecker - Managing Partner and Global Head of Wipro FullStride Cloud.

With over two decades of industry experience, Jo is known for his expertise in leading complex transformation projects for large global organizations and for driving consistent growth for the businesses he leads. Most recently, Jo was the Global Head of Cloud Infrastructure Services.