Most Java-based organizations use Java for AI development – report

Azul Systems, in its Azul 2025 State of Java Survey & Report, says that Java leads Python and JavaScript for AI development in Java shops, based on a survey of 2,039 Java professionals on six continents.

The report, released January 28, finds that 50% of organizations that use or deploy Java-based applications and infrastructure use Java to code AI functionality and Java-based libraries to build AI functionality. The survey found that 50% of these companies use Java for coding AI functionality, followed by JavaScript at 44%, and Python at 41%.

Java developers are actively leveraging AI to drive innovation, enhance application functionality, and deliver significant business value, the report states. Also, 72% of survey participants said compute consumption will have to grow for them to support Java applications with AI functionality.

Azul’s report also takes a swing at Java rival Oracle.

“Two years after Oracle announced its employee-based Java pricing and licensing, concern over Oracle’s pricing model remains high. Organizations are responding in kind, with 88% considering migrating away from Oracle to another Java distribution/provider,” according to the report.

In the cloud realm, Azul found that nearly two-thirds of organizations that run Java workloads in the cloud said Java accounts for more than 50% of cloud compute costs, while 71% of those organizations said they have more than 20% unused, but purchased, cloud capacity. Organizations are addressing cloud costs through strategies such as implementing new internal rules for managing cloud instances, adopting finops practices, upgrading to more efficient compute instances and processors, and leveraging high-performance Java Development Kits (JDKs).

Azul’s survey of 2,039 Java professionals from six continents asked how enterprises are dealing with Oracle Java pricing and licensing, what strategies they’re adopting to address cloud costs, about factors impacting their devops productivity, and about Java’s role in their AI development.

Other findings in the Azul 2025 State of Java Survey & Report:

Groovy was the most popular Java-based language used by participants in the survey, snagging 36% of respondents, followed by Scala at 35%, Kotlin at 33%, and traditional Java at 28%.

62% of participants said dead or unused code affects devops productivity somewhat or to a great extent, while 33% said 51% or more of their devops team’s time is wasted addressing false positives from Java-related security vulnerabilities.

68% of respondents report that more than 50% of their applications are Java-built or run on a JVM.

 Java 17 was the most commonly used Java version for production applications (34% of organizations), followed by Java 21 (31%), Java 11 (23%), and Java 8 (23%).

Spring Boot was the most commonly used microservices framework for Java-based applications (42% of organizations), but was closely followed by Micronaut (39%), DropWizard (37%), Quarkus (31%), and Vert.x (25%). Only 7% of participants said they don’t use a Java-based microservices framework.

 Almost all Java-based companies — 85% of survey participants — pay for Java support.