Java Modernizes: New Tools for AI and Quantum Age

Ever since Sun Microsystems open sourced Java in 2006, innovation has flowed out of its loyal and creative development community. Current Java corporate caretaker Oracle has maintained the tradition it inherited in the 2010 acquisition of Sun with the release of Java 24 (Oracle JDK 24) at JavaOne 2025 earlier this month.

The latest iteration of the widely used programming language and development platform contains significant advancements aimed at developer productivity and improved enterprise-grade application capabilities. JDK 24 includes an upgrade that spans language features, libraries, security, tooling and performance — just about everything important in the platform.

According to IDC analyst Arnal Dayaratna, Java continues to evolve and remain valuable by addressing the needs of developers, particularly in the realm of AI-powered applications. Georges Saab, senior vice president at Oracle, cited the inclusive nature of Java 24, which encompasses more than 20 new features, including AI and post-quantum cryptography support. This release reinforces Oracle’s commitment to a predictable six-month release cadence.

Several quantum-related items are listed below. The most important AI-related upgrade is JEP 489 (Vector API), which accelerates vector computations — crucial for AI inference and compute-intensive tasks. That one undoubtedly will get a lot of use from AI developers.

Here, according to Oracle specifications, are details on JDK24 improvements:

Security Strengthening

Java 24 prioritizes security, particularly in the face of emerging quantum computing threats.

• JEP 478 (Key Derivation Function API) improves cryptographic security for data in transit.
• JEP 496 (Quantum-Resistant Module-Lattice-Based Key Encapsulation Mechanism) and JEP 497 (Quantum-Resistant Module-Lattice-Based Digital Signature Algorithm) provide implementations of quantum-resistant mechanisms, an important step toward post-quantum cryptography support. These features address the need for secure communication and data authentication in a post-quantum world.

Language Updates

Key language features aimed at streamlining development and improving code reliability are.

• JEP 488 (Primitive Types in Patterns, instanceof, and switch) provides greater uniformity and expressiveness by extending pattern matching to primitive types, benefiting AI inferencing applications.
• JEP 492 (Flexible Constructor Bodies) enhances code reliability by introducing distinct prologue and epilogue phases in constructor bodies, simplifying logic placement.
• JEP 494 (Module Import Declarations) simplifies the reuse of modular libraries, particularly beneficial for beginners and developers integrating AI logic.
• JEP 495 (Simple Source Files and Instance Main Methods) offers a smooth learning curve for new programmers and allows experienced developers to write concise small programs.

Library Additions

Several significant library improvements.

• JEP 485 (Stream Gatherers) enhances the Stream API, allowing for custom intermediate operations and more efficient data transformation.
• JEP 484 (Class-File API) provides a standard API for parsing, generating and transforming Java class files.
• JEP 487 (Scoped Values) improves the sharing of immutable data across threads, enhancing performance and robustness.
• JEP 489 (Vector API) accelerates vector computations, crucial for AI inference and compute-intensive tasks.
• JEP 499 (Structured Concurrency) simplifies multithreaded programming, improving maintainability and reliability.

Tooling, Performance Optimizations

• JEP 493 (Linking Run-Time Images without JMODs) enables the creation of custom run-time images without JMOD files, reducing JDK size.
• JEP 450 (Compact Object Headers) reduces object header sizes in the HotSpot JVM, improving heap size and performance.
• JEP 475 (Late Barrier Extension for G1) optimizes the G1 garbage collector, enhancing efficiency and code quality.
• JEP 483 (Ahead-of-Time Class Loading and Linking) improves application startup time.
• JEP 490 (ZGC: Remove the Non-Generational Mode) simplifies ZGC maintenance.
• JEP 491 (Synchronize Virtual Threads without Pinning) enhances virtual thread scalability.
• JEP 404 (Generational Shenandoah) introduces experimental generational collection capabilities.
• JEP 479 (Remove the Windows 32-bit x86 Port) and JEP 501 (Deprecate the 32-bit x86 Port for Removal) streamline the JDK build and test infrastructure.

Removing Certain Features for Security Purposes

The OpenJDK community also emphasized the removal of what it considers unsafe features, including JEP 472, JEP 486 and JEP 498, to maintain Java’s integrity and align with best practices. JEPs 472, 486 and 498 are part of a broader effort to enhance Java platform integrity by default, focusing on restricting potentially unsafe features and practices.

Here’s why these JEPs are considered unsafe:

JEP 472: Prepare to Restrict the Use of JNI: JNI (Java Native Interface) allows Java code to interact with native (C/C++) code, which can introduce security risks and portability issues. The Security Manager was a mechanism to restrict permissions for remotely loaded code (like applets), but it has become less relevant with the decline of applets.

JEP 486: Permanently Disable the Security Manager: JEP 486 permanently disables the Security Manager, removing a feature that was causing friction and complexity. This move simplifies the platform and reduces the potential for security vulnerabilities associated with legacy security mechanisms.

JEP 498: Warn upon Use of Memory-Access Methods in sun.misc.Unsafe: JEP 498 issues warnings when unsafe methods are used, and these methods are deprecated and will be removed in future releases. This prepares developers for the eventual removal of these unsafe APIs and encourages them to use safer alternatives.

Cloud Integration and Community Support

As one would expect, Oracle claims that Java 24 offers improved performance and cost savings when run on Oracle Cloud Infrastructure (OCI) because they have been co-engineered. The Oracle Java Universal SE Subscription provides new support, the company said, including the Java SE Subscription Enterprise Performance Pack and access to GraalVM.

At the JavaOne event, industry experts weighed in on JDK 24. Frank Greco of Crossroads Technologies highlighted the Vector API’s enhancement for AI applications, while Richard Fichtner of XDEV Software appreciated the Stream Gatherers for efficient data transformations. Dr. Venkat Subramaniam of Agile Developer, Inc. praised the Stream Gatherers, Scoped Values and Structured Concurrency. CodeRanch moderator Jeanne Boyarsky noted the benefits of flexible constructors and the potential of Stream Gatherers. JetBrains’ Marit van Dijk emphasized her company’s commitment to providing day-one support for Java 24 in IntelliJ IDEA.