Oleksii - stock.adobe.com
APAC firms ditch Oracle Java to cut costs amid AI demands
Rising Oracle Java licensing fees and the massive infrastructure requirements of AI workloads are driving the region’s businesses to migrate to OpenJDK distributions
A large majority of technology professionals in Southeast Asia are moving their companies away from Oracle’s Java software, driven by rising licensing costs and the need to free up budget for power-hungry artificial intelligence (AI) workloads.
According to the 2026 State of Java survey and report, released by Java infrastructure software firm Azul, 97% of Java professionals in the region are concerned about rising costs resulting from Oracle’s shift to employee-based pricing, a figure significantly higher than the global average of 92%.
As a result, 78% of Southeast Asian enterprises, particularly those in highly regulated sectors such as financial services and utilities, have already migrated, or plan to migrate, their Java infrastructure software to non-Oracle OpenJDK, the open source version of the Java Development Kit for the Java Platform, Standard Edition (Java SE).
Java is a programming language that has served as the backbone of enterprise software for decades, used by banks, retailers and governments to run critical applications. While the language is free, the specific distribution provided by Oracle has seen changes in pricing models and audit practices that have unsettled businesses.
Speaking at a media briefing on the findings, Dean Vaughan, Azul’s vice-president for Asia-Pacific, said the move to non-Oracle OpenJDK distributions is also being fuelled by the massive infrastructure requirements of artificial intelligence (AI) workloads.
While the public focus is often on consumer-facing tools such as ChatGPT, he explained that the backend infrastructure that powers AI tools is mostly built on Java. These include the Kafka event streaming platform, the Elasticsearch search and analytics engine, and the HBase NoSQL database, all of which are Java-based.
But as companies look to optimise their AI infrastructure, simply cutting idle virtual machines is no longer enough, said Vaughan. The focus is now on re-architecting the AI infrastructure stack with high-performance Java runtimes to reduce compute requirements. “If we can replace the standard Java with an optimised Java that requires less compute to deliver the same AI service, then we can help companies save a lot of money on their cloud or infrastructure expenditure,” he said.
‘White glove’ service
However, Vaughan said the likes of Amazon Web Services (AWS) only provide a standard OpenJDK distribution and don’t offer bespoke optimisation services. By contrast, Azul offers a “white glove” service where its engineers work with clients to manually tune the Java software. This involves analysing logs and adjusting memory configurations to squeeze more performance out of existing hardware.
“We could agree upfront on a 10–15% performance improvement on your application,” he said. “If we give you a 15% performance improvement, then you’ll need 15% less compute. That’s 15% fewer virtual machines, storage and virtual LAN.”
Vaughan said this lets Azul negotiate success-based pricing with customers, where it takes a cut of the money the client saves on their cloud or hardware bills in the case of on-premise deployments, particularly for organisations in Southeast Asia, where data sovereignty laws may require them to operate workloads in private datacentres.
“If they could just replace the Java layer with a faster Java, then they get that speed and benefit without having to buy the hardware,” he added, citing a payment processing client in the region that used this method to manage rapid growth when they could not purchase physical servers fast enough.
Payara acquisition
To further support companies in modernising their Java infrastructure, Azul recently acquired Payara, an open source company with deep expertise in enterprise-grade Java (Jakarta EE) and open source application servers.
The acquisition also adds Payara Micro, a lightweight Jakarta EE runtime optimised for containers and microservices, to Azul’s portfolio. Vaughan noted that the product is particularly attractive to organisations that need to scale their services quickly without being weighed down by heavy, traditional software infrastructure.
The report also found distinct differences between the Australia and New Zealand (ANZ) and Southeast Asian markets. ANZ firms are moving more aggressively, with 87% migrating, or planning to migrate at least part of their Oracle Java deployments to a non-Oracle OpenJDK distribution, while a significant 67% intend to or have already migrated their entire Java estate.
Despite the momentum towards Java optimisation, the report indicates there’s still work to do. In both ANZ and Southeast Asia, roughly two-thirds of respondents reported having over 20% unused compute capacity in their public cloud environments, reflecting ongoing inefficiencies in how Java workloads are provisioned.
In ANZ, security noise also adds another layer of complexity, with 49% of enterprises now dealing with Java-related common vulnerabilities and exposures on a daily or weekly basis – a sharp rise from 33% in 2025.
In addition, 38% of ANZ respondents said that between 26–50% of their team’s time is wasted on security vulnerability false positives related to Java Virtual Machine-based workloads, and 17% say more than half of their team’s time is wasted on chasing false positives.
Read more about Java
- Oracle Java licensing explained – addressing complexity, cost and audits: Now that Oracle Java SE is sold as a subscription, organisations using the Oracle JDK have a minefield of licences to navigate.
- Tips for migrating to OpenJDK: Interest in OpenJDK, and in commercial support for it, has intensified as Oracle Java SE becomes increasingly expensive.
- James Gosling invented Java in 1995. It has been kept up to date and now supports artificial intelligence, but Oracle Java is no longer the preferred choice.
- Java Community Process chair Heather VanCura talks up efforts to keep Java up to speed with the needs of developers in the era of cloud and agile software development.
