Organizations already generate and store vast amounts of data. However, data volumes are expected to skyrocket due to AI adoption.
The dramatic increase is largely due to the need to retain data for longer periods to fuel AI applications. Ninety percent of organizations that are using AI believe that retaining data longer improves the success of AI initiatives.
Both on-premises and cloud storage are expected to grow to accommodate increased data volumes. However, organizations expect to store 65 percent of their data in the cloud versus just 35 percent in-house. By 2028, 69 percent of data will be stored in the cloud.
Budget overruns and business delays could be the result. In a separate study conducted by Vanson Bourne, 62 percent of organizations said they exceeded their cloud storage budgets in 2024. One-fourth said the budget overruns were substantial.
Data egress and usage fees were the primary culprits. Cloud providers charge for retrieving and transferring stored data. These fees vary based on the provider, storage class, region and other factors such as data access frequency and transfer volume. As organizations access data stored in the cloud for AI applications, they’re seeing egress and usage fees skyrocket.
These fees aren’t just blowing a hole in organizations’ cloud budgets. More than half (56 percent) said they have delayed IT projects and business initiatives due to high data access costs.
Many organizations move data to the cloud to reduce the need to invest in onsite infrastructure. Cloud storage platforms allow organizations to offload the costs and administrative overhead that come with supporting physical hardware. Some industry studies have claimed that in-house storage can cost five times more per gigabyte to own and operate than cloud storage.
However, organizations are often disappointed if cost reduction is their primary goal. In addition to usage and egress fees, organizations pay more per gigabyte to store data in a “hot” storage tier. Designed for data that needs to be accessed quickly and frequently, this premium tier is optimized for high performance and low latency, making it suitable for real-time analytics and AI.
Cloud providers also charge for certain storage operations, such as reading or writing a file or moving data between tiers. These fees are typically quite small but can add up with applications that perform operations on large numbers of storage objects. Organizations will also pay more for storage and bandwidth if they replicate their data across multiple cloud regions.
As organizations continue to adopt AI, they will need to find ways to rein in cloud costs. Here are three steps they can take to gain better control over their cloud storage budgets.
Re-Evaluate Storage Tiers. AI applications may need ready access to data that would have been archived in the past. Cloud providers often charge higher egress fees from cold storage tiers. It may make sense to move data to higher storage tiers to reduce egress costs even though the cost per gigabyte of capacity is higher.
Optimize Egress Fees. There is a range of techniques organizations can use to manage egress costs. Using private networking or dedicated bandwidth connections can reduce the cost of large data transfers, and batching data extractions into larger downloads can reduce the number of individual egress transactions. Consolidating workloads within the same region can minimize inter-region transfers.
Use FinOps and Cost Modeling. FinOps is a financial management discipline for managing opex costs across the enterprise. When applied to cloud spending, it enables organizations to better track their cloud resources and impose accountability for cloud usage. Cost modeling allows them to predict costs based on data growth rates and select the best storage platforms and tiers.
Technologent’s cloud and storage experts can help you leverage these and other strategies for controlling cloud storage costs. Our team can evaluate your environment, index and assess unstructured data, and help you select the right cloud providers and storage models to optimize your budget.