Java developers are simultaneously abandoning Oracle’s distribution to cut costs while expanding their use of the language to support production-grade AI.
The 2026 State of Java Survey and Report, published by Azul and based on responses from over 2,000 Java professionals, identifies a defensive transition away from unpredictable licensing fees alongside an offensive move to operationalise AI workloads.
The migration accelerates
Migration from Oracle’s distribution has gained significant momentum. 81 percent of organisations have migrated, are migrating, or plan to migrate at least some of their Oracle Java estate to non-Oracle OpenJDK distributions. 63 percent of respondents intend to migrate their entire Java estate.
Cost drives these decisions for 37 percent of respondents. Since the 2023 introduction of employee-based pricing, dissatisfaction has mounted. The report notes that 92 percent of Java professionals are concerned about Oracle’s pricing. Conversely, only seven percent describe themselves as “not at all concerned,” a figure that has nearly halved since the previous year.
Operational risk also influences the exodus. 21 percent of organisations report being subjected to an Oracle Java audit. This exposure, coupled with uncertainty regarding future licensing changes, drives 26 percent of migration decisions. Another 31 percent of enterprises cite a preference for open-source alternatives as their primary motivation.
High-performance Java for production AI
While Python dominates model training, Java is establishing itself as the engine for AI in production. 62 percent of organisations now use Java to code AI functionality, an increase from 50 percent last year.
The adoption reflects a need to integrate machine learning models into established enterprise architectures. As companies move from experimentation to deployment, requirements shift toward scalability and security. 31 percent of respondents state that more than half of the applications they build now contain AI functionality, supported by libraries such as JavaML and the Deep Java Library (DJL).
To support this transition, technical leaders are prioritising specific runtime capabilities. 35 percent of participants identify long-term support (LTS) for modern Java versions as essential for an AI-enabled landscape, alongside built-in security (34%) and observability insights (32%).
Cloud waste and optimisation
Cloud cost management remains a priority, with 97 percent of organisations taking action to reduce public cloud spend. Despite these efforts, inefficiencies persist. 74 percent of organisations report having more than 20 percent unused compute capacity in their cloud environments. This buffer is often provisioned to mitigate slow start-up times or unpredictable runtime behaviour.
In response, 41 percent of engineering teams are adopting high-performance Java platforms specifically to lower compute costs. Among enterprises heavily invested in the language (those where at least 90% of applications run on Java) adoption of high-performance platforms to improve application output rises to 81 percent.
The report identifies friction points slowing DevOps teams. “Dead” or unused code affects productivity for 63 percent of respondents. This technical debt creates a reluctance to refactor, as teams hesitate to remove legacy components, slowing cycles and increasing risk.
Security alerts further complicate operations. 56 percent of enterprises deal with Java-related Common Vulnerabilities and Exposures (CVEs) on a daily or weekly basis, a rise from 41 percent in 2025. Much of this effort appears misdirected: 30 percent of teams spend more than half their time investigating false positives, often triggered by scanners flagging libraries that are present in the codebase but never executed in production.
“Java continues to prove its durability and strategic importance as enterprises navigate one of the most transformative periods in modern computing,” said Scott Sellers, Co-Founder and CEO at Azul. “From powering the next generation of AI-driven applications to helping organisations regain control of cloud spend and modernise their estates, Java remains at the centre of innovation and operational excellence.”
For Java developers, the data suggests a need to clear operational overhead – such as unnecessary licensing fees, unused cloud capacity, and dead code – to accommodate the AI workloads expanding into production environments.
See also: Agile Manifesto at 25: How it reshaped software delivery
Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events. Click here for more information.
Developer is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.