Asia Pacific’s AI ambitions are colliding with the past. Outdated infrastructure is quietly sabotaging the region’s artificial intelligence race, and a widening revenue gap is exposing who is falling behind.
Key takeaways
- Legacy infrastructure is the single biggest barrier to AI adoption across Asia Pacific, with nearly half of organisations unable to build new applications without major modernisation first.
- The revenue gap between digital leaders and the rest is not theoretical: leaders generate 71% of revenue from digital products while mainstream peers manage just 23%.
- Companies that ignore their technical debt are on borrowed time, with IDC forecasting a 50% higher AI failure rate for laggards by 2027.
Across the Asia Pacific, boardrooms are buzzing with AI ambitions. But beneath the optimism, a stubborn obstacle is stalling progress: the creaking weight of legacy technology that companies have long deferred modernising.
New research commissioned by MongoDB and conducted by IDC paints a sobering picture. A survey of 1,400 organisations across eight markets found that 43% reported their existing architecture makes it impossible to build new applications without extensive modernisation, systems that their own staff describe as too rigid, too costly, and too slow for what the AI era demands.
The findings land at a pivotal moment. Companies across the region have moved from experimenting with AI pilots to attempting full-scale production deployments. That transition is brutally exposing the gap between ambition and infrastructure.
The data problem no one wants to talk about
At the root of the crisis is data quality. The most commonly cited software development challenge was data management and poor-quality data, named by 32% of organisations. Close behind were outdated database technology that cannot support AI workloads and the difficulty of embedding security into development without sacrificing speed, each cited by 31% of respondents.
In plain terms, many companies are attempting to build next-generation AI systems on platforms designed for a previous era of computing.
Supporting new AI initiatives was the main reason for modernising databases and applications, with 46% of organisations naming it as their top driver. Yet the path to modernisation is proving treacherous. Nine in ten organisations surveyed reported having experienced failed modernisation initiatives, with siloed and poor-quality data identified as the main obstacle.
A revenue divide is opening up
The research does more than catalogue frustration. It identifies a consequence that finance leaders cannot ignore: a measurable and growing commercial gap between companies that have modernised and those that have not.
A smaller group of companies described as leaders are pulling away from their peers, generating 71% of revenue from digital products and services, compared with just 23% among mainstream peers. Those leading organisations share a common trait. 58% are running multiple programmes to reduce legacy constraints and build cloud-ready foundations for AI systems in production, treating modernisation not as a project with an end date but as a permanent discipline.
The cost of standing still is rising sharply. IDC has forecast that organisations that do not address technical debt will face a 50% higher failure rate and rising costs for AI initiatives by 2027.
What the experts are saying
IDC’s Senior Research Director for Asia Pacific, Dr William Lee, was direct about what the data reveals. He described high-quality, integrated data as the essential fuel that determines the accuracy and performance of an AI application, and said many organisations are being held back by rigid legacy architectures that lack the flexibility and scalability to handle the high volume of unstructured data required for AI.
MongoDB’s Managing Director of CXO Advisory, Thorsten Walther, framed the issue in board-level terms, arguing that AI has made technical debt an urgent priority for senior leadership and that the research shows strategic modernisation unlocks AI opportunities and supports significant revenue growth.
A real-world example
The study points to Bendigo Bank as a concrete illustration of what modernisation can achieve. The bank moved a core banking application away from legacy relational database technology to MongoDB Atlas and used AI-assisted tools to break the work into smaller releases, reducing development time by up to 90% and cutting costs to one-tenth of a traditional migration, all without service outages.
The path forward
IDC outlined a set of steps for organisations seeking to improve their AI readiness, including stronger data quality and governance, modernising architectures that slow application development, building cloud-ready hybrid operating models, and investing in skills and change management.
The survey covered organisations with at least 100 employees across Australia, China, Hong Kong, India, Indonesia, Singapore, South Korea, and Thailand, spanning developers, IT decision-makers, and senior executives.
The picture that emerges is of a region at a crossroads. Those who treat modernisation as a strategic priority are pulling ahead commercially. Those who continue to defer it are not simply falling behind on technology benchmarks. They are falling behind on revenue, resilience, and their ability to compete in an economy that AI is rapidly reshaping.
