Delta CEO puts power and skills at heart of Thailand’s AI push

Thailand needs stronger power infrastructure, a larger pool of AI talent and deeper local innovation to turn rising investment in artificial intelligence and electrification into lasting economic value, according to Victor Cheng, CEO of Delta Electronics (Thailand).

Speaking at the Delta Future Industry Summit 2026 on Friday (August 14), Cheng described AI and electrification as two major opportunities developing simultaneously, while warning that their rapid expansion would put increasing pressure on electricity supply, networks and industrial resources.

Demand for AI computing is roughly doubling each year, he noted, while electrification is accelerating across electric vehicles, buildings, infrastructure and factory automation. Supporting both trends will require energy that is reliable, efficient and, where possible, clean and renewable.

Cheng estimated that hyperscale operators would invest more than US$600 billion this year, with the pace of expansion already putting considerable strain on supply chains.

Lead times for some commodities have reached 40 weeks or more, while demand is increasing not only for integrated circuits but also for the materials and infrastructure needed to support data centres and server racks.

The pressure extends beyond chips. Cheng identified electricity, cooling, optical systems, networking and storage as areas where bottlenecks could emerge, but also as fields in which Thailand could capitalise on its existing manufacturing strength.

“Thailand already has a strong footprint in manufacturing capacity involved in many of these products across the board,” he noted, citing a June World Bank publication that he said ranked South Korea, Taiwan, Malaysia and Thailand among the strongest exporters of equipment supporting the AI industry.

Power becomes the main bottleneck

Power supply has become the leading concern for AI infrastructure developers, according to Cheng. He projected that electricity demand would roughly double between 2024 and 2030 to about 950 terawatt‑hours.

Thailand has close to three gigawatt-hours of data-centre projects in the pipeline, he added, while access to power remains one of the main constraints. About 40% of announced capacity has been delayed because of power interconnection issues rather than difficulties in constructing the hardware itself.

Hyperscale operators therefore have to look beyond building data centres and increasingly manage their own energy requirements through power purchase agreements or, in some cases, their own power-generation infrastructure.

Cheng argued that stronger “grid-to-edge” infrastructure could help meet rapidly rising energy demand. Microgrids, in particular, could strengthen energy resilience and security while making electricity supply more cost-effective.

Smart energy management and infrastructure could therefore turn current power bottlenecks into new business and commercial opportunities, he added.

Government and private sector must work together

Technology and business alone cannot solve the challenge, Cheng stressed. The government must provide the policy and implementation framework, while the private sector contributes technology, solutions, capital and other forms of investment.

He pointed to microgrids, battery storage, on-site solar generation and EV charging as examples of infrastructure that could help create a more resilient energy system capable of supporting rapid growth.

But removing the power bottleneck would only be the first step. Thailand must also ensure that incoming investment generates lasting domestic value through skills development, innovation and economic models that allow more of the benefits to remain in the country.

Building AI skills and developing the workforce is therefore urgent, Cheng emphasised. Thailand needs people who can support the industry, attract further investment and help retain more of the resulting value within the economy.

AI investment must create wider economic value

Cheng sees the pressures created by AI expansion and rising energy demand as an opportunity for Thailand to establish itself as a Southeast Asian AI hub and attract further investment.

He pointed to Thailand’s National Semiconductor Roadmap for 2025–2050, which aims to move the country beyond assembly and further into chip design and fabrication.

Foreign manufacturing investment, combined with training programmes for graduates, could help accelerate the semiconductor ecosystem and prepare more workers to participate in the industry, he noted.

Smart manufacturing could also help Thailand move towards higher-value, automation-driven production and advance further along industrial value chains.

The potential benefits of AI, Cheng added, should extend well beyond data centres. Applied AI could support established areas of the Thai economy including tourism, agriculture, healthcare and beauty care, helping strengthen industries that already have deep domestic roots.

Speed will be critical because many countries are competing for the same investment and industrial opportunities, he stressed. Thailand will therefore need to move quickly and remain agile if it wants to capture the benefits of the current shift.

Ultimately, Cheng argued, the goal should not be limited to constructing AI data centres. Thailand should apply AI across a broader range of sectors, retain more skills and economic value domestically and build a sustainable ecosystem in which investment, talent and economic growth reinforce one another over the long term.