Storage is structural cost and below I take a look at why AI is increasing storage costs and how you can manage it.
Structural Costs > Storage
AI needs faster access to much larger data sets, which increases storage costs.
Storage is where your information is kept and you can store data in a variety of ways, including databases and file systems. How quickly you need to access the data is an important factor in understanding storage costs: the faster you need to retrieve your information, the higher the cost. Hot storage is accessible in sub-milliseconds, warm storage in seconds and cold storage in minutes or hours.
To manage cloud storage costs, content experts work to decide which type of storage system to use to retain what type of information and what to destroy. These decisions have a large impact on the storage portion of the monthly cloud bill, which is typically relatively stable month to month.
How is AI changing things?
AI uses the highest performing, and most expensive, storage mediums. These storage types include high performance vector databases and parallel file systems. While traditional storage was described by its capacity (gigabytes GB, terabytes TB) the types of storage used by AI are described based on throughput, which includes how fast information can be retrieved.
AI also consumes storage at a much larger scale and with higher variability, which is already causing significant storage cost overruns in enterprises worldwide. Managing these highly variable storage costs is complex and can be challenging as organizations shift from managing passive costs with low variability to shifting to larger and more highly variable storage costs.
What can you do about it?
To control the amount and variability of storage costs incurred by AI, you can:
- Use information management strategies to ensure only the best, most relevant data is in active use
- Begin monitoring new dimensions that indicate AI storage consumption, including type of storage, number of queries and amount of storage caused by reindexing
- Deploy agentic AI thoughtfully to ensure the higher number of queries it produces are applied where they can have the most value
In Summary: Ensure your AI has access to the leanest, most relevant information to control storage cost.
Hi
Support this Work
You can purchase the book on Amazon:
You can listen to the podcasts for free through Patreon, Apple Podcasts or YouTube Music: