The latest Long-Term Support (LTS) version of the Java Development Kit, JDK 25, is now available, bringing with it a wide range of capabilities designed to modernize application development, boost performance, and even streamline artificial intelligence workflows. This release introduces 18 new features, with Oracle highlighting five as particularly beneficial for AI-related use cases.

A New Chapter in Java’s LTS Cycle

JDK 25 follows March 2025’s JDK 24, which was a short-term release. As an LTS version, JDK 25 will benefit from at least eight years of commercial support. The last major LTS milestone was JDK 21, unveiled in September 2023. Developers can download the latest toolkit at jdk.java.net.

Features with AI in Mind

Several of the new enhancements are aimed at making Java a stronger foundation for AI-driven applications:

• Primitive types in patterns and switch statements simplify the integration of numerical operations within business logic.
• Structured concurrency makes it easier to manage workloads involving multiple parallel tasks—common in machine learning.
• Scoped values improve efficiency when sharing immutable data across threads.
• The Vector API accelerates inference and computational workloads.
• Module import declarations simplify incorporating AI frameworks, libraries, or services.

Together, these updates promise smoother integration between Java codebases and advanced AI workflows.

Profiling and Performance Enhancements

Developers working on Linux will gain a notable boost through JDK Flight Recorder’s (JFR) CPU-time profiling. By leveraging the Linux kernel’s CPU timers, JFR can now capture more accurate performance metrics, helping teams fine-tune their applications. This addition joins two other JFR-related improvements in JDK 25: cooperative sampling and method timing and tracing.

Performance-focused changes don’t stop there:

• Ahead-of-time (AOT) method profiling reduces warmup time by making past execution profiles available as soon as the JVM starts.
• AOT command-line ergonomics simplify creating caches that speed application startup.
• Generational Shenandoah GC is now a fully supported feature after proving its stability and performance in demanding workloads.

Security and Cryptography

JDK 25 also makes strides in cryptography with a preview of PEM encoding support. This new API offers developers a straightforward way to encode and decode cryptographic keys, certificates, and revocation lists in the widely used PEM format. Compatibility with existing standards like PKCS#8 and X.509 ensures smoother interoperability across systems.

Smaller but Impactful Changes

Additional updates include compact object headers, flexible constructor bodies, stable values API, and the removal of the 32-bit x86 port, reflecting Java’s ongoing evolution toward efficiency and modernization.

Why It Matters

For developers, the impact of JDK 25 will be felt across performance optimization, security, and next-generation workloads. Whether the goal is reducing application startup time, tightening cryptographic operations, or harnessing Java for AI, this release signals that the language continues to adapt to the demands of modern computing.