Organizations have long struggled with cloud budget overruns. AI is making the problem worse.
Cloud storage costs have skyrocketed thanks to AI adoption. According to a recent survey conducted by Vanson Bourne, 62 percent of organizations exceeded their cloud storage budgets in 2024, up from 53 percent in 2023. Data egress, access and usage fees are the primary culprits. These fees account for 49 percent of cloud storage costs.
Of course, data storage costs are also increasing. Organizations already store vast amounts of data, and are now retaining more data for longer periods to support their AI initiatives. In a recent survey conducted by Recon Analytics, 61 percent of organizations said they expect their storage volumes to double, with the typical large firm storing 300 petabytes by the end of 2026.
Most of that data is destined for cloud storage. Organizations expect that 69 percent of their data will be stored in the cloud by 2028, with just 31 percent remaining on premises.
Understanding Cloud Storage Tiers
Cloud storage is relatively inexpensive, and can be a real bargain compared to buying and maintaining onsite hardware. According to some industry studies, in-house storage can cost five times more than cloud storage when factoring in both capital costs and administrative overhead.
A lot depends, of course, on the storage tier. Cloud storage is divided into tiers according to data access frequency. These tiers range from hot (frequent, immediate access like live production data) to cool (less frequent access like semi-active content) to cold (rarely accessed data for long-term retention). Cold storage is cheapest, until the data needs to be accessed.
Organizations typically choose cold storage for archival data, but AI is causing those plans to change. AI may give rise to a need to access that data, and organizations may be hit with retrieval fees and per-operation fees as well as standard data egress charges.
Turning Up the Heat
Because cold storage is inexpensive, it has slow retrieval times, and accessing data from cold storage can degrade application performance. In the Vanson Bourne survey, 84 percent of organizations said they access data stored in cold tiers on a weekly or monthly basis, and 98 percent said they struggle with poor data performance and data access penalties.
Hot storage delivers the high performance and low latency needed for real-time analytics and AI, but it costs significantly more per gigabyte. Organizations may also incur charges for moving data between tiers and for removing data from cold storage before a minimum duration. These costs add up — 56 percent of organizations said they delayed business or IT initiatives due to unexpected cloud storage costs.
Tips for Controlling Cloud Storage Costs
Luckily, there are steps organizations can take to rein in cloud storage costs and get a handle on their budgets. Organizations can reduce egress fees and gain higher speeds and improved security by using dedicated private connections that bypass the public Internet. Organizations can also consolidate downloads to reduce the number of egress transactions, and use compression to reduce the volume of data that must be transferred.
Organizations may need to rethink their cloud storage tiering strategies and consider migrating complete applications to the cloud to reduce the need for frequent data egress. Caching frequently accessed data in an optimal tier close to the application can reduce egress costs as much as 50 percent by minimizing the need to repeatedly fetch data across the network.
Cerium’s cloud and AI experts can help you leverage these and other techniques for controlling cloud storage costs. We can also help you use cloud-native management tools to gain insight into data transfer patterns and optimize data flows. If cloud storage is taking up an ever-larger chunk of your IT budget, give us a call to discuss the various options.
