Companies often set their artificial intelligence priorities at a regional level. The systems, investment and operating conditions supporting those priorities are far more local.
An AI use case that is commercially viable in one Asia-Pacific (APAC) market may face different costs, regulations, infrastructure constraints and customer expectations in another. Scaling AI across the region therefore involves more than applying the same technology plan market by market.
IDC expects AI and generative AI spending in APAC, including China and Japan, to increase fivefold from US$73 billion in 2024 to US$370 billion by 2029. Infrastructure provisioning is expected to remain the largest use case, accounting for approximately 39 percent of that spending.
The scale of investment makes readiness more important. When an AI project remains a contained experiment, its limitations may be manageable. Once it supports customer service, employee workflows or operational decisions, the expectations change. The organization must be confident that the application can perform consistently, use the right information and recover when something goes wrong.
AI workloads also change what organizations require from their storage and data environments. More information may need to be retained, accessed and processed across different systems and locations. As applications become more important to daily operations, organizations also need clearer expectations around cost, availability and recovery.
At this point, AI becomes more than an innovation project. It becomes a business infrastructure decision.
The business expectations change at scale
A production system has to work repeatedly, not only during a demonstration. It may need to serve different teams and markets, use information held across several parts of the organization and meet existing expectations around security, compliance and business continuity.
Hitachi Vantara’s 2025 State of Data Infrastructure research illustrates the size of this operational challenge. Across six surveyed APAC markets, 73 percent of respondents said their organizations were already using AI extensively or regarded it as critical. At the same time, 78 percent said the complexity of their data infrastructure was increasing rapidly or extremely rapidly, while 96 percent expected to need external help in at least one area.
A practical example from outside the region shows what can happen when AI becomes part of an operating process. At Hitachi Vantara’s manufacturing facility in Norman, Oklahoma, the company had to manage approximately 3,000 product variants amid shifting demand and increasingly customized requirements.
AI was applied to demand forecasting, customer proposals, inventory planning and manufacturing. Some applications required inventory and supplier information from around the world to be consolidated, while others retrieved and integrated information held across multiple systems. The reported results included a 19 percent improvement in demand-forecast accuracy, a 50 percent reduction in global inventory and an 84 percent reduction in the lead time for configure-to-order production.
The significance lies not only in the numbers. The applications addressed defined operating problems and depended on information, domain knowledge and processes working together. That is a different proposition from introducing an isolated AI tool and waiting to see where it might add value.
Readiness is not the same in every market
The same implementation cannot simply be copied across APAC. The region’s markets vary in size, geography, regulation, energy availability, established technology investments and access to specialist skills.
Hitachi Vantara’s research found that 89 percent of respondents in India described AI as being used extensively or considered critical to their organization. Yet 54 percent identified a shortage of skilled personnel as a challenge, 46 percent cited regulatory requirements and 43 percent pointed to integration with established systems.
Singapore presented a different combination. Sixty-six percent reported extensive or critical AI use, while security was identified as a challenge by 61 percent, followed by skills and compliance. In Taiwan, reported AI use was lower at 43 percent, with skills and cost among the more prominent concerns.
These figures should not be used to rank markets. They show that organizations can share the same AI ambition while encountering different obstacles.
In a market where power, land or physical capacity is constrained, organizations may place greater emphasis on efficiency and making better use of existing investments. In a rapidly growing market, skills, integration and the ability to expand without disrupting current operations may matter more. Where regulation or data-location requirements are prominent, organizations may need to decide more carefully where information is stored and who can access it.
Common goals do not require identical implementation
Regional organizations still benefit from consistency. Common governance, business objectives and expectations around resilience can prevent fragmented decision-making. But consistency should not mean requiring every market to follow an identical implementation plan.
The starting point should be the business outcome. Leaders need to understand which process the AI application will support, who will depend on it and what level of interruption the organization can accept. This provides a clearer basis for determining the investment and operational support required.
The organization should also examine the information the application will rely on. Information may be spread across business units, systems and countries, with different owners and restrictions. If those conditions are not understood early, they can affect the cost, timetable and usefulness of the project later.
Cost should be considered across the full life of the project. A successful pilot can increase usage quickly, but demand forecasts, models and operating requirements can also change. Investment should therefore be phased where possible, with room to expand or adjust rather than committing every market to a fixed capacity assumption from the outset.
Resilience should reflect the application’s business importance. An experimental internal tool and an application supporting customer transactions do not carry the same consequences if they are interrupted. Expectations around availability, protection and recovery should increase as the organization becomes more dependent on the application.
A regional leadership responsibility
The shift from AI experimentation to dependable business use requires closer cooperation between leadership, technology, operations and local market teams.
Regional leaders have an important role in defining the common outcome, setting expectations and ensuring that investment decisions support the business strategy. Local teams can then identify the conditions that may affect implementation, including regulation, infrastructure availability, market maturity and customer expectations.
AI readiness cannot be judged only by the sophistication of a model or the success of a pilot. It also depends on whether the organization can support the application consistently, manage its information responsibly and sustain the investment as usage grows.
For companies operating across APAC, the AI ambition may be regional, but many of the decisions that make it viable will be local. The strongest strategies will combine common outcomes and governance with enough flexibility to reflect how each market actually operates.
Wendy Koh is Vice President and General Manager of Asia-Pacific at Hitachi Vantara, where she leads the company’s regional strategy and works with customers and partners on data infrastructure and hybrid cloud adoption. Based in Singapore, she has more than 30 years of experience in IT consulting, infrastructure and digital transformation. Before joining Hitachi Vantara, she was Executive Vice President for Southeast Asia at Capgemini and held leadership roles at Cisco, Juniper Networks and NetApp. Koh also serves as an independent non-executive director at ASMPT Limited.
Editor’s note: This contributed article has been lightly edited for clarity, concision and TNGlobal house style. The substance of the author’s contribution has been preserved.
Share your perspective: TNGlobal welcomes contributed insights and expert commentary from across Asia’s technology and innovation ecosystem. Submit a contribution for editorial consideration, or explore more conversations in our TNGlobal INSIDER and TNGlobal Q&A and Interviews archive.
Featured image: Road Ahead on Unsplash
AI is now a trusted insider in your business. Is it being governed like one?