Do I Need a Database Anymore?
by Ginger Grant | February 2, 2026
This sounds like a provocative question. It's a fair question for any developer architecting a new system in 2025. After all, databases have been the default choice for decades. When I ask organizations why they're keeping so much data, I hear similar refrains:
- “AI might be able to gain insight from that data so we are keeping it.”
- “Legal wants us to make sure we save all the data.”
- “Storage is cheap so why get rid of data?”
Storage is cheap, especially cloud storage, which offers physically redundant, secure storage for about $28 a terabyte per month. Databases on the other hand are not cheap, as with databases there is a need to maintain the structure and perhaps pay a license fee depending on the type of database, and probably have a number of people working on tuning it, creating backup jobs of various kinds. For years, databases were the only way non-technical users could retrieve organizational data.
How Have Data Storage and Retrieval Technology Improvements Impacted Databases?
In the last fifteen years, there have been a number of advancements in storage compression and memory interface techniques. Previously, data was loaded from disk to memory. Having the fastest disk you could buy would make it perform better. You need to continually monitor performance and look at the common queries to add indexes to improve performance. You need to pay for licenses based upon the amount of data you want to store, and the number of cores used on the server. Now the data is compressed and stored in memory and much of the monitoring uses algorithms so the tuning requirements go away. AI relies on data. The same chips and data centers that are powering advancements in AI are also being used to store and retrieve data for more mundane purposes, such as querying data. You no longer need to store data in a proprietary storage container to be able to organize and retrieve data. Large volumes of data are being stored in newer open-source formats designed to compress and sort data rapidly on disks and in memory using familiar query languages and tools already known to millions of people. These tools are not as expensive to maintain or store data in as they have been in the past. Microsoft Fabric provides methods for storing and retrieving vast amounts of data for a fixed price.
Do I Need a Cloud Database Anymore or Can I Just Use Fabric?
Fabric has incorporated newer open-source technology in the way it stores data while accommodating existing methods, so that users don’t have to choose a single solution. There are four different options for storing your data inside of Fabric: Lakehouse, Warehouse, Eventhouse and SQL DB. Given that you can create all of these inside of Fabric, which one is best suited to the task at hand? Well, if you want to fully embrace data lakes and use native Spark structures, you should use a lakehouse. When you want to combine skills you have honed, such as doing SQL Server data warehousing to load and query your data primarily with T-SQL, then warehousing would be the option you would select. Eventhouses were created to monitor and store streaming real-time data. If you have a database and want to continue to use it, that option is provided for you with SQL DB. But what if you already have data stored in a database in the cloud? Should you look to add to the data that you already have there? Well, if you are looking at the cost of storage and retrieval and the ease of retrieval and use, I think the answer is no. Selecting a Fabric Data store vs an Azure Data storage element will most of the time be a lower cost option.
How Can Fabric Handle Large Amounts of Data?
What about large amounts of data? Don’t you need a big database for that? No. You can store data in files and still be able to use the data for analyzing questions using the SQL language. Now in Fabric you will be using data stored on open source Parquet files in an ever-expanding cloud, allowing you to have terabytes of data in each table.
Looking to the Future
Changes in technology have impacted many different aspects of our lives from where we buy goods and services to where we go for entertainment. Databases are not immune to these changes either. As storage and memory improvements continue to be made, it makes sense to move to places where the advantages are available now with open-source technologies, and that place is Microsoft Fabric.