As AI becomes more accessible to the average person, it becomes critical to build a robust architecture that supports AI solutions. to this end,Microsofthas updated its .NET and JDBC ecosystem to natively support vector data types, ensuring a high-performance environment for AI workloads.

Speaking of .NET, the Sqlvector class was added in Microsoft.Data.SqlClient 6.1.0. This class is able to handle data in a vector format rather than the inefficient JSON array format used previously. In comparison, vector types provide up to 50x better read performance, up to 3.3x better write performance, and up to 19x better bulk copy performance. These performance tests were conducted on SQL Server 2025 Preview, with vector columns up to 1998 characters and 10,000 records processed per operation.

Additionally, since JSON serialization is no longer required, lengthy string representations are not required and memory consumption is lower. Finally, 32-bit floating point vectors are currently natively supported, but may be expanded to support other numeric data types in the future.

Switch to a JDBC environment, where there is a new VECTOR data type for the JDBC driver for SQL Server 13.1.0. It works directly with inserts, selects, stored procedures and bulk copying. This release also removes string-based vector processing, making it ideal for integration with Java-powered AI applications, including those that provide semantic search.

Notably, these improvements are available across platforms, including SQL Server 2025 (17.x) preview, Azure SQL Database, Azure SQL Managed Instance, and Microsoft SQL Database Fabric preview. It's worth mentioning that these improvements are achieved through optimization of the Tabular Data Stream (TDS) protocol, so they are only visible in TDS 7.4 or later.

Clients that do not support this protocol will continue to use the varchar(max) data type and treat the data as a JSON array for backward compatibility, which means they will not experience any of the above enhancements.

For all the details, check out the release notes: https://github.com/dotnet/SqlClient/releases/tag/v6.1.0

JDBC Release Notes: https://github.com/microsoft/mssql-jdbc/releases/tag/v13.1.0