Asia Pacific (Apac) governments are positioning the region to benefit from the global artificial intelligence (AI) boom, rolling out infrastructure investment, targeted fiscal support and industrial policies aimed at lifting productivity. However, a widening gap is emerging between policy ambition and corporate capability as most companies remain stuck at early stages of AI use despite rising adoption.
The Mastercard Economics Institute projects that AI adoption and targeted fiscal support will serve as meaningful tailwinds for Apac's growth in 2026. South Korea, Japan, India and Hong Kong are leading the region in both corporate and consumer uptake of AI tools. Governments across the region are backing this momentum with industrial policies targeting AI hubs, data centres, smart cities and semiconductor development.
Singapore offers a clear view of both the opportunity and the constraint. Amazon Web Services’ (AWS) Unlocking Singapore’s AI Potential 2025 report reveals that 48% of businesses in the city-state now use AI, up from 40% a year earlier.
Yet 65% of these AI-adopting businesses remain focused primarily on basic applications, such as publicly available chatbots and scheduling assistants, rather than transformative uses, such as custom AI systems or research and development.
“2023 was about AI experimentation, 2024 about moving to production, and the coming years will bring truly industrial-scale AI implementation,” Priscilla Chong, country manager for AWS Singapore, tells The Edge Singapore.
A two-tier AI economy
See also: University of Nottingham Malaysia rolls out AI agent for student recruitment
The gap between adoption and impact has created, as the AWS report describes, a two-tier AI economy, especially in Singapore.
Despite higher adoption rates, larger enterprises often struggle to scale AI beyond isolated use cases. About 62% of large firms in Singapore use AI, but only 30% have a comprehensive AI strategy that outlines how the technology supports their core business objectives.
In contrast, start-ups are more likely to embed AI into their operating models, with nearly half placing AI at the centre of their business proposition.
See also: AI in mental health: Friend or foe?
“For companies looking to scale beyond basic adoption, three critical areas demand attention. First, building consolidated ‘AI factories’ underpinned by a sound data strategy; second, developing governance frameworks that balance innovation with appropriate guardrails; and third, cultivating the specialised technical capabilities needed for extensible AI systems that can work autonomously across diverse workflows,” advises Chong.
Skills, not technology, are the bottleneck
The most significant brake on progress is talent. The same AWS study reveals that 43% of the businesses in Singapore cite a lack of digital skills as the main barrier to expanding AI use.
“What’s unique about the AI skills challenge is that it’s not static—skills evolve with technology. The conversation about skills isn’t new, but what’s changed dramatically is how AI fluency looks different for every role,” says Chong.
She continues: "Consider what a marketer will need two years from now—not just prompt engineering, but the ability to evaluate AI-generated content against brand values, integrate multimodal outputs across channels, and leverage customer behaviour predictions ethically. It's about bridging domain knowledge with role-specific AI skillsets."
The skills gap is particularly pronounced among small and medium-sized enterprises (SMEs). AI literacy skills per employee at firms with 11 to 50 workers grew 67% year-on-year, lagging the 99% at companies with more than 1,000 employees, according to LinkedIn’s Work Change Special Report: How Small Businesses Can Win in 2026.
Training also lags at smaller firms, with less than half of employees in SMEs receiving employer-provided AI training. Despite limited support, workers are teaching themselves and using AI independently for everyday tasks like writing emails and summarising notes (45%) as well as for complex strategy and data analysis (26%).
To stay ahead of the latest tech trends, click here for DigitalEdge Section
Specialisation model
With advanced talent scarce and unevenly distributed, the AWS report argues that Singapore is moving toward a “specialisation model” in which firms increasingly plug into external AI expertise rather than build everything internally.
Chong explains: “Singapore's ecosystem is developing a natural specialisation that creates a powerful innovation cycle. Start-ups build new AI products and business models, enterprises prove these solutions can scale, and the public sector acts as a 'trust multiplier' that encourages wider adoption. Each segment reinforces each other."
Singapore-based Hypotenuse AI exemplifies this approach. The start-up helps retailers accelerate time to market by using AI to generate SEO-optimised product descriptions, enrich product data, and edit product images in days instead of months. It serves major e-commerce brands, including Quiksilver and Billabong, across the US, Europe and Asia in over 30 languages by building its solution on AWS's cloud infrastructure.
"Retail actually has tremendous untapped potential given the wealth of data it already captures—from consumer purchasing patterns and browsing behaviour to inventory movements and seasonal trends. This rich data foundation positions retail perfectly for advanced AI implementation," says Chong.
At the enterprise level, Keppel built KAI, its internal AI platform powering specialised agents for research, investment analysis and real-time news monitoring. Built on Amazon Bedrock, the platform integrates multiple generative AI models (including Anthropic Claude and Amazon Nova) to allow teams to choose the best model for each use case while embedding governance and security controls from the outset.
Public-sector leadership remains central to this ecosystem. According to the AWS study, 71% of businesses say they are more likely to adopt and expand their use of AI when the public sector does so, while 75% of start-ups say government adoption is crucial to their ability to scale.
Commenting on that finding, Chong says: "I see this as a trust multiplier effect. When the public sector demonstrates successful AI implementation, it gives businesses confidence to adopt similar technologies. This creates a powerful ripple across industries."
