The talent and infrastructure shaping Asia’s AI build-out
Most conversations about artificial intelligence begin with its eventual impact: productivity, automation, new products and the jobs that may change with them.
But before AI can transform businesses at scale, something more fundamental has to happen.
The infrastructure has to be built.
That means computing capacity, semiconductors, data centres, electricity, cooling and networks. It also means research capability, cybersecurity, governance, regulated testing environments and ultimately the people capable of designing, operating and commercialising the ecosystem.
Across Asia, that build-out is already underway.
The more immediate talent question may therefore be less about which occupations AI eventually replaces, and more about which capabilities are needed before AI can scale in the first place.
AI may be digital, but much of its foundation is physical
Generative AI has made computing demand much more visible.
Data centres require substantial power and cooling. Advanced models depend on high-performance chips, cloud infrastructure and connectivity. Enterprises need engineers capable of integrating AI into existing systems, while regulated sectors need governance frameworks before applications can move from experimentation into production.
The result is a value chain stretching well beyond software development.
Capital must finance the infrastructure. Engineers must build it. Technology specialists must operate it. Risk and governance professionals must make its use safe and compliant. Product and commercial teams must ultimately turn the technology into something businesses and consumers will pay for.
That is why Asia’s AI investment cycle is beginning to create a much broader talent story.
Singapore: expanding compute while anchoring capability
Singapore provides a useful example of how physical infrastructure and higher-value activity can develop together.
Under its Green Data Centre Roadmap, Singapore is targeting at least 300MW of additional data-centre capacity in the near term. Its second Data Centre Call for Application makes at least 200MW available, with applicants assessed not simply on capacity but also on contributions to innovation, talent development, ecosystem partnerships and sustainability.
That distinction matters.
Singapore is not treating data-centre development solely as a real-estate or infrastructure exercise. The intention is for computing capacity to contribute to a wider digital ecosystem.
There are already signs of that activity around the infrastructure.
Google has invested around US$5 billion in technical infrastructure across four data centres and cloud regions in Singapore, while expanding its local AI R&D footprint and maintaining a Google DeepMind research presence.
AI company Cognition selected Singapore as its Asia-Pacific headquarters, with regional leadership, operations and growth based in-market.
Revolut, meanwhile, is building high-skilled Singapore teams across engineering, product, data and AI. The company doubled its local headcount in 2025 and plans to triple its workforce over the following three years as Singapore anchors its Asia expansion.
These examples do not mean every data centre creates an equivalent number of high-value jobs. But they illustrate a broader point: compute capacity becomes more economically significant when research, product development and regional decision-making grow around it.
Hong Kong: connecting capital, research and regulated adoption
Hong Kong is assembling a different combination of capital, computing infrastructure, research and financial-services adoption.
One capital lever is the HK$10 billion Innovation and Technology Industry-Oriented Fund, designed to use government capital to attract additional private investment into strategic sectors.
The fund covers five thematic areas: life and health technology; AI and robotics; semiconductors and smart devices; digitalisation and transformation; and future and sustainable development. Professional fund managers establish sub-funds, with the combined target size across them set at at least HK$40 billion.
Alongside this capital initiative, Hong Kong is expanding its computing and research infrastructure.
The Government says Hong Kong’s overall computing power has reached around 5,000 petaFLOPS, while the Sandy Ridge data-facility cluster is intended to provide additional capacity for AI-related data and computing requirements. Sixteen InnoHK laboratories already specialise in AI and robotics, while a new Hong Kong AI Research and Development Institute is intended to support research and the transformation of R&D outcomes.
The financial-services sector offers another layer: GenA.I. Sandbox++.
Rather than simply encouraging financial institutions to adopt AI, Hong Kong’s financial regulators are providing a supervised testing environment. Participating firms can receive targeted regulatory guidance, technical support and access to GPU computing resources at Cyberport’s AI Supercomputing Centre.
The first cohort includes 36 use cases from 30 financial institutions working with 27 technology partners, with particular emphasis on agentic AI applications.
This matters because regulated adoption requires a different capability mix from pure AI development: technology risk, model governance, compliance, cybersecurity and business-domain expertise all become part of the equation.
The build-out is increasingly distributed across Asia
Singapore and Hong Kong are only two parts of a much wider regional story.
Malaysia has rapidly expanded its role in hyperscale data-centre infrastructure. In April 2026, AirTrunk announced another US$3 billion for two Johor facilities, taking its planned Malaysian capacity beyond 700MW.
Vietnam, meanwhile, continues to deepen its semiconductor and advanced-manufacturing position. Samsung is planning around US$1.5 billion for a semiconductor testing facility north of Hanoi, adding to Vietnam’s established role in semiconductor assembly and testing.
The same pattern extends beyond these markets. India, for example, is seeing substantial investment in cloud and AI infrastructure: Amazon announced an additional US$13 billion through 2030, while major global and domestic technology companies are building data-centre and AI capacity across the country.
The important point is not that one market “owns” a particular part of the AI stack.
It is that different concentrations of capability are emerging across Asia, and increasingly depend on one another.
The talent requirement extends far beyond AI engineering
For recruitment and workforce planning, this is where the infrastructure story becomes particularly relevant.
AI build-out can create demand across capital and investment roles such as infrastructure investing, project finance and asset management; physical-infrastructure disciplines including engineering, power, construction, networks and data centres; technology capabilities spanning cloud, semiconductors, cybersecurity and AI engineering; governance disciplines such as model risk, compliance, technology risk and data governance; and commercial roles responsible for turning capability into usable products and services.
This does not mean every market will require the same mix.
Singapore’s combination may place relatively greater emphasis on AI R&D, regional product leadership and regulated technology. Malaysia’s data-centre expansion can deepen engineering, power and infrastructure capability. Vietnam’s semiconductor investment reinforces manufacturing and technical engineering skills. Hong Kong’s regulated financial adoption can create greater demand at the intersection of technology, governance and financial services.
The talent story is therefore regional as well.
Talent could become part of the bottleneck
Physical capacity is not the only constraint.
Data-centre expansion is increasingly testing power grids, water supply and cooling requirements. Large infrastructure commitments also require sufficient utilisation to justify their economics.
Talent presents a different problem: specialist capability cannot simply be switched on when a new facility opens or a new investment programme is announced.
Engineering expertise, cloud architecture, cybersecurity, model governance and technology-risk capability take time to develop. So do experienced leaders capable of bringing together technology, regulation and commercial strategy.
For hiring teams, that creates a planning question well before large-scale AI adoption arrives:
Which capabilities will we need to build, operate, govern and commercialise the infrastructure surrounding AI — and how long will those talent pipelines take to develop?
Looking ahead
We may not yet know which AI applications will ultimately create the most economic value.
But the infrastructure being built around them is already reshaping where capital, technology and talent are being deployed across Asia.
The opportunity therefore may be less about predicting the eventual AI winner, and more about recognising the capabilities taking shape while that future is still being built.
For organisations, that means looking beyond immediate AI hiring and understanding the wider ecosystem required to support adoption.
For professionals, it means some of the most consequential AI-related opportunities may emerge in roles that do not carry “AI” in their title at all.
References:
- IMDA Singapore
- EDB, Singapore
- Hong Kong Government Budget
- Hong Kong ITF Government
- Securities and Futures Commission, Hong Kong – OFC Regime and fund re-domiciliation