Java News Roundup: TornadoVM, A2A Java, Open Liberty, Open J Proxy, Gradle, EclipseStore

The Evolution of the JDK: Progress on Build 23

The Java Development Kit (JDK) 28 continues its steady progression toward release, with the arrival of Build 23 this week. As the primary reference implementation for the Java programming language, the JDK 28 early-access cycle is being closely monitored by enterprise teams and infrastructure providers alike. Build 23, following the refinements introduced in its predecessor, focuses primarily on stability and the resolution of long-standing technical debt.

Historically, the early-access phase of a new JDK version serves as a critical feedback loop between the OpenJDK community and the broader developer base. By utilizing the Java Bug Database, developers can influence the final stability of the release. The current trajectory for JDK 28 suggests a focus on memory management and JIT (Just-In-Time) compiler efficiency, which remains the cornerstone of Java’s competitive advantage in high-throughput environments.

Accelerating Performance: TornadoVM and TornadoInsight

One of the most notable technical milestones this week is the release of TornadoVM 8.1.0. TornadoVM has established itself as a critical bridge for developers looking to offload compute-intensive Java tasks to heterogeneous hardware, including GPUs, FPGAs, and multi-core CPUs. The 8.1.0 update introduces a significant architectural shift: the ability for a kernel to perform in-place read/write operations on device allocations owned by external libraries. By bypassing the need for expensive memory copying or the insertion of custom array headers, TornadoVM significantly lowers the latency barrier for high-performance computing (HPC) applications.

Complementing this, the Beehive Lab has released TornadoInsight 2.1.0, an IntelliJ IDEA plugin that acts as a vital tool for developers attempting to navigate the complexities of hardware-accelerated Java. The introduction of experimental AI-assisted transformation marks a shift toward automating the translation of standard Java methods into optimized GPU kernels. This feature lowers the entry threshold for developers who may lack specialized experience in OpenCL or PTX programming, potentially accelerating the adoption of AI-driven computing within traditional enterprise Java stacks.

Modernizing Agentic Workflows with A2A SDK

The A2A (Agent-to-Agent) Protocol has gained traction as a standard for building decentralized agentic applications. With the release of version 1.5.0 of the A2A Java SDK, the project is focusing on observability and resource optimization. The inclusion of the OpenTelemetryPushNotificationSenderDecorator is particularly noteworthy; it aligns agentic communication with the broader OpenTelemetry ecosystem, allowing for standardized tracing of agent interactions.

The ability to filter database queries through authorized JPA task listings is a strategic improvement. By reducing scan costs and minimizing deserialization overhead, the SDK directly addresses the performance bottlenecks often associated with large-scale agentic systems. This development highlights the ongoing maturation of agent-based software architectures, moving from experimental prototypes to robust, production-ready infrastructures.

Quarkus and Open Liberty: The Cloud-Native Front

Quarkus, the Kubernetes-native Java stack, has reached its third beta milestone for version 4.0.0. The ecosystem continues to prioritize compatibility with emerging technologies, as evidenced by the new support for the Valkey data store within the Redis Client extension. Valkey, a high-performance alternative to Redis, has seen rapid adoption, and its inclusion in Quarkus signals the framework’s commitment to staying agile in the face of shifting infrastructure trends. Additionally, the new @CheckedTemplateParam annotation enhances the developer experience in Qute, the template engine, by simplifying parameter management without requiring verbose compiler flags.

Parallel to the rapid pace of Quarkus, Open Liberty has released version 26.0.0.11, a General Availability (GA) update. As a production-hardened platform, Open Liberty’s latest release focuses on security and lifecycle management. The introduction of the quiesceTimeout configuration is a vital feature for microservices architectures, allowing developers to fine-tune the shutdown process of server instances to ensure that active requests are handled gracefully before the container terminates. This level of granular control is essential for maintaining high availability in distributed environments.

Infrastructure and Tooling: Gradle and Open J Proxy

Gradle 9.9.0 introduces substantial improvements to the build lifecycle. The move to enable incremental compilation for Kotlin DSL scripts using the Kotlin Build Tools API is a performance-oriented update that will save countless hours in large-scale CI/CD pipelines. Furthermore, the expansion of the Problems API provides a more structured approach to diagnostic reporting, helping teams identify and resolve build-time conflicts more rapidly.

In the realm of interoperability, Open J Proxy 1.1.0 (Release Candidate 3) continues to bridge the gap between Java and C++. By improving support for PostgreSQL and MariaDB within the C++ ODBC client and streamlining the cross-platform build process, Open J Proxy ensures that legacy data architectures can be integrated into modern Java workflows with minimal friction.

Data Management: EclipseStore and Micrometer

The EclipseStore 4.3.0 release underscores the importance of efficient object-to-storage mapping. The refactoring of the VectorIndex interface for JVector Index is a response to the increasing demand for high-concurrency, large-scale data structures. By improving crash safety and load performance, EclipseStore maintains its position as a high-performance persistence layer for Java objects. The transition to a lightweight VanillaJS storage viewer further modernizes the management interface, shedding the weight of older Vaadin-based GUI components.

Micrometer, the industry-standard metrics facade, has also seen progress with the second release candidates for both Metrics (1.18.0) and Tracing (1.8.0). The support for fractional updates in Atlas counters addresses a precision issue that had previously limited the accuracy of high-resolution telemetry. Meanwhile, the efficiency gains in the Tracing component—specifically regarding OpenTelemetry baggage allocation—demonstrate the project’s ongoing focus on reducing the overhead of observability tools in high-traffic applications.

The Road Ahead: JReleaser 2.0

Andres Almiray’s announcement of the JReleaser 2.0 roadmap sets a clear trajectory for the coming months. JReleaser has become an indispensable part of the Java release lifecycle, automating the process of publishing software to various repositories and package managers. The transition to a 2.0 architecture will include breaking changes designed to simplify the configuration model and expand support for newer release distribution platforms. The developer community is currently preparing for these changes, with a 2.0.0-SNAPSHOT now available for early testing.

Broader Implications and Industry Analysis

The developments of this week illustrate a clear, unified direction for the Java ecosystem. Three distinct themes emerge from these updates:

1. AI and Hardware Acceleration: The push toward AI-assisted development and hardware-agnostic computing (via TornadoVM and AI-integrated tooling) suggests that Java is aggressively positioning itself as a primary language for the next generation of high-performance AI applications.
2. Developer Experience and Performance: Whether through Gradle’s incremental build improvements, the reduction of boilerplate in Quarkus, or the memory-efficient tracing in Micrometer, the community is obsessed with optimizing the "inner loop" of development. This is essential for organizations that need to iterate rapidly on complex, cloud-native systems.
3. Observability as a First-Class Citizen: The integration of OpenTelemetry across multiple SDKs and frameworks—from the A2A SDK to Micrometer—highlights the shift toward standardized, cross-platform diagnostic data. In modern distributed systems, the ability to trace a transaction from an agent to a database is no longer a luxury; it is a fundamental operational requirement.

As we look toward the remainder of 2026, the convergence of these technologies suggests a more modular, high-performance, and observable Java ecosystem. The stability provided by Open Liberty and the experimental agility of Quarkus demonstrate that the platform continues to cater to a diverse set of needs, from strictly governed enterprise environments to cutting-edge cloud-native startups. Developers are encouraged to review the provided documentation and release notes for these updates, as many include critical security patches and performance enhancements that will benefit production deployments immediately.