Java Ecosystem Evolution: JDK 27 Features, AI Frameworks, and Developer Tools

The Java ecosystem continues its steady evolution with several significant developments across the language, frameworks, and tooling. This week's highlights include important Java Enhancement Proposals (JEPs) targeting JDK 27, milestone releases for major frameworks, and new SDKs for AI and agent-based applications. These developments reflect Java's ongoing adaptation to modern development paradigms while maintaining its commitment to stability and performance.

OpenJDK Evolution: JDK 27 JEPs

The OpenJDK project continues to refine features for JDK 27, with several JEPs advancing through the review process. These enhancements demonstrate Java's ongoing commitment to improving developer productivity and expanding the language's capabilities.

JEP 532: Primitive Types in Patterns, instanceof, and switch

This JEP has progressed to "Targeted" status for JDK 27, representing its fifth preview iteration after being introduced in JDK 23. The feature extends pattern matching capabilities by allowing primitive types in all pattern contexts, thereby enhancing the instanceof and switch constructs to work with all primitive types.

The significance of this enhancement lies in its ability to simplify code that operates on primitive types. Previously, developers needed to handle primitive types differently from reference types in pattern matching contexts. This JEP bridges that gap, enabling more consistent and expressive code across different data types.

JEP 533: Structured Concurrency

Elevated from "Candidate" to "Proposed" status, this JEP introduces structured concurrency to Java as a seventh preview feature. Structured concurrency treats groups of related tasks running in different threads as a single unit of work, thereby streamlining error handling and cancellation, improving reliability, and enhancing observability.

This approach represents a fundamental shift in how concurrent programming is approached in Java. By treating related tasks as a single unit, developers can avoid common pitfalls of concurrent programming such as thread leaks, inconsistent error handling, and complex cleanup logic. The feature builds on previous incubation periods in JDK 19 and 20, and preview iterations in JDK 21 through 26.

JEP 531: Lazy Constants

Formerly known as StableValues, this JEP has reached "Targeted" status for JDK 27 as a third preview feature with some notable changes. The modifications include:

1. Removal of the isInitialized() and orElse() methods from the LazyConstant interface
2. Addition of a new factory method ofLazy() that can create stable, pre-defined elements for all three Java collection types (List, Set, and Map)

This feature aims to provide a more efficient way to handle constants that are expensive to compute but may never be used. By making constants lazy, Java applications can reduce startup time and memory usage, particularly beneficial in large applications and containerized environments.

JEP 536: JFR In-Process Data Redaction

This JEP has advanced to "Candidate" status, proposing to enhance the JDK Flight Recorder (JFR) to redact sensitive information before the JFR completes its recording. The data that may be redacted includes command-line arguments, initial values of environment variables, and system properties.

The importance of this feature cannot be overstated in today's security-conscious environment. By enabling in-process redaction, Java applications can maintain detailed observability through JFR without exposing sensitive information that could lead to security vulnerabilities. This represents a significant step forward in making observability tools more security-aware.

Framework Updates: Spring AI, GlassFish, and Quarkus

Several Java frameworks released significant updates this week, reflecting the ecosystem's adaptation to AI/ML workloads and cloud-native patterns.

Spring AI Milestone 5

The fifth milestone release of Spring AI 2.0.0 introduces several enhancements that improve the framework's capabilities for AI application development. Notable additions include:

• Custom implementations of the StructuredOutputConverter interface that can now participate in the native structured output pipeline, providing improved flexibility for output formatting
• A new method isConversationHistoryEnabled() added to the ToolCallAdvisor.Builder class for better conversation management

Spring AI continues to position itself as a comprehensive framework for integrating AI capabilities into Spring applications, with this release focusing on improving the developer experience and flexibility of the framework's AI integration patterns.

GlassFish 9.0.0 Milestone 2

The second milestone release of Eclipse GlassFish 9.0.0 brings several updates to this open-source Jakarta EE application server:

• A minimal version of JDK 21 with support for JDK 25
• Implementations of Jakarta Security 5.0-M2, Jakarta Faces 5.0-M2, and Jakarta Contexts and Dependency Injection 5.0-M1 specifications
• A new implementation of the JSON Web Token (JWT) specification with Jakarta Security and Eclipse Soteria

This release demonstrates the ongoing evolution of Jakarta EE application servers, with a focus on supporting newer Java versions and implementing the latest specifications. The JWT implementation is particularly noteworthy as it addresses a common requirement for modern web applications.

Quarkus 3.35.0

Quarkus continues to innovate with its release of version 3.35.0, introducing:

• An experimental build-time dependency "tree shaker" that identifies and excludes unused classes from runtime dependencies to reduce application JAR sizes
• A new Reactive Transactions extension that enables the Spring Framework @Transactional annotation to support Hibernate Reactive

The tree shaker feature addresses a common challenge in containerized environments where application size directly impacts startup time and resource consumption. By reducing the size of the runtime dependencies, Quarkus applications can achieve faster startup times and smaller memory footprints, which is particularly beneficial in serverless and microservice architectures.

Developer Tooling Updates

The Java development tooling ecosystem saw several significant releases that improve build processes, dependency management, and release automation.

Gradle 9.5.0

The general availability release of Gradle 9.5.0 introduces notable improvements to build diagnostics and task management:

• Enhanced Gradle Task failure diagnostics and reporting that now include provenance information and clearer logging when the client JVM is incompatible
• A new disallowChanges() method added to the DomainObjectCollection interface that prevents elements from being added to or removed from the collection

These improvements address common pain points in Gradle builds, particularly around debugging build failures and managing build configuration integrity. The provenance information in task failure diagnostics will help developers more quickly identify the root causes of build issues.

JReleaser 1.24.0

JReleaser, a utility that streamlines creating project releases, has been updated to version 1.24.0 with several new features:

• Support for Zernio added to the list of announcers
• A reproducible flag to configure when JReleaser should output consistent content

These enhancements expand JReleaser's capabilities for release automation, particularly for projects that require reproducible builds. The Zernio integration provides another channel for announcing releases, while the reproducible flag ensures that release artifacts can be consistently generated across different environments.

AI and Agent Development SDKs

The Java ecosystem is increasingly focusing on AI and agent-based applications, with several new SDKs released to support these workloads.

Agent2Agent Java SDK 1.0.0 Beta 1

The first beta release of Agent2Agent (A2A) Java SDK 1.0.0 introduces a Java library that implements the Agent2Agent Protocol for running agentic applications as A2AServers. Developed as a collaboration between the Quarkus and WildFly teams, this SDK provides:

• Support for JSON-RPC, gRPC, and HTTP+JSON/REST transport protocols
• Use of the JdkA2AHttpClient class as a default for more flexible HTTP client support
• Structured error codes with details

This SDK represents an important step toward standardizing how agent-based applications communicate in Java. By supporting multiple transport protocols, it provides flexibility for different deployment scenarios while maintaining a consistent programming model.

Google Agent Development Kit 1.2.0

Google's Agent Development Kit (ADK) for Java has been updated to version 1.2.0 with several enhancements:

• New custom telemetry headers added to the Google Cloud BigQuery interface
• The McpSessionManager class has been added to the builder defined in the McpAsyncToolset class

This release continues Google's investment in making Java a first-class language for developing AI agents on Google Cloud platforms. The telemetry enhancements improve observability for agent applications, while the McpSessionManager class provides better session management capabilities.

LangChain4j 1.14.0

The formal release of LangChain4j 1.14.0 introduces several new features for building applications with large language models:

• A new attribute IMMEDIATE_IF_LAST added to the ReturnBehavior enum that returns an execution loop result from an AI service when this tool is positioned last in the LLM response
• A new OpenAiOfficialResponsesChatModel class to align with the existing OpenAiOfficialResponsesStreamingChatModel class for use with the OpenAI Responses API

These enhancements improve the flexibility of LangChain4j when working with LLMs, particularly in scenarios where tool execution order matters. The IMMEDIATE_IF_LAST behavior provides more control over when results are returned, which can be useful in complex multi-step reasoning scenarios.

Looking Ahead: Micronaut 5.0

The Micronaut Foundation has provided a preview of the upcoming Micronaut Framework 5.0 release, scheduled for 2Q 2026. This version will feature:

• JDK 25 as a baseline
• Support for cloud vendors that have migrated to JDK 25
• Performance improvements
• Updated code examples in the Micronaut Guides and documentation

Additional planned features include support for Kotlin 2.3, Groovy 5.0, and Jackson 3.0, as well as adoption of JSpecify. Notably, Micronaut 5.0 will drop support for RxJava 2, reflecting the broader Java ecosystem's shift toward more modern reactive programming approaches.

Conclusion

The Java ecosystem continues to evolve in response to emerging technologies and development patterns. The JDK 27 JEPs demonstrate Java's commitment to improving developer experience and expanding language capabilities, particularly in concurrent programming and pattern matching. Framework updates show Java's adaptation to AI/ML workloads and cloud-native patterns, while new SDKs highlight Java's growing role in agent-based applications.

These developments collectively indicate a Java ecosystem that is both maintaining its core strengths while actively embracing new paradigms. As Java continues to evolve, developers can expect a language and ecosystem that remains relevant to modern development challenges while preserving the stability and performance characteristics that have made Java a dominant force in enterprise software development.