Thailand gains ground in regional AI supply-chain race

Thailand has secured a place in several critical parts of the global artificial-intelligence supply chain, with its electronics industry producing equipment used in data-centre power, cooling, optical communications, networking, servers and data storage.

The country does not manufacture the most advanced processors used to train AI models. Its factories, however, supply or assemble much of the infrastructure needed to connect, power and cool those chips and store the data they process.

Companies with relevant operations in Thailand include Delta Electronics, Fabrinet, Celestica, Cal-Comp Electronics, Western Digital, Seagate Technology, Lumentum, Coherent, Quanta Computer and Schneider Electric. Their activities place Thailand across several sections of the AI hardware chain rather than tying it to a single product.

The expansion of AI computing has increased demand not only for semiconductors but also for high-density electrical systems, liquid cooling, optical components, network switches, server assembly and storage equipment.

Power and cooling support denser computing

AI servers require large amounts of electricity and produce considerably more heat than conventional computing systems, increasing demand for power-conversion and thermal-management equipment.

Delta Electronics supplies power and cooling products for data centres, including equipment designed for high-density servers and liquid-cooled installations.

Schneider Electric operates in the same broader market through data-centre power management, automation and cooling systems.

These operations give Thailand a role in the infrastructure surrounding advanced processors, even though the country does not yet produce the chips themselves.

The rapid growth of data-centre investment is also placing greater demands on Thailand’s electricity grid, water supply and access to cleaner energy.

The Thailand Board of Investment has increased its scrutiny of proposed data-centre projects, including their electricity consumption and environmental impact.

Investment applications submitted to the BOI reached about 1.47 trillion baht in the first half of 2026, with digital infrastructure accounting for approximately 1.12 trillion baht.

The agency received applications for 36 data-centre projects worth more than 728 billion baht in 2025. It approved another seven data-centre and hosting projects worth about 96.88 billion baht in early 2026.

The projects could increase domestic demand for electrical systems, cooling equipment, servers, network products and storage hardware made or assembled in Thailand.

Optics and networks connect AI systems

Large AI clusters must transfer vast quantities of information between processors, servers and data-centre buildings. This makes optical communications and high-capacity networking another important part of the supply chain.

Fabrinet operates advanced optical-manufacturing facilities in Thailand. Lumentum and Coherent specialise in photonics, lasers and communications technology used in data infrastructure.

Celestica’s operations in Thailand add electronics engineering, network-equipment assembly and supply-chain services. Cal-Comp Electronics and Quanta Computer provide further manufacturing capacity connected to computing and server equipment.

Not every product made by these companies in Thailand is necessarily used in an AI system. Their existing capabilities nevertheless give the country a production base that can respond to rising demand from cloud and data-centre operators.

Western Digital and Seagate Technology also anchor Thailand’s longstanding hard-disk-drive industry.

Storage receives less attention than graphics processors, but AI systems require substantial capacity for training data, archives, applications and generated content. Hard drives remain part of that infrastructure alongside solid-state storage.

Thailand therefore participates in several connected parts of the physical AI supply chain, covering power, cooling, optical equipment, networking, assembly and storage.

Regional rivals pursue higher-value investment

Thailand’s position is developing alongside stronger competition from Malaysia, Singapore and Vietnam.

Malaysia has an established semiconductor industry covering integrated-circuit design, assembly, packaging and testing. It also produces printed circuit boards and storage components while expanding its data-centre market and investment in advanced engineering.

Singapore competes through semiconductor fabrication, research and development, manufacturing equipment and regional headquarters. Its skilled workforce and research base allow it to attract activities carrying greater technological and intellectual-property value.

Vietnam has expanded rapidly as an electronics-manufacturing centre and continues to offer incentives for investment in semiconductors, electronic components and other high-technology industries.

Thailand’s advantages include established industrial estates, automotive and electronics supply chains, export infrastructure and a large base of multinational manufacturers.

Much of its current role, however, remains concentrated in foreign-owned manufacturing, assembly and contract production. Product design, advanced engineering, research and intellectual property generally generate higher returns than manufacturing alone.

Thailand’s semiconductor strategy focuses on fields linked to its existing industrial base, including power semiconductors, sensors, photonics, discrete devices and analogue chips.

The long-term plan targets more than 2.5 trillion baht in investment and the development of 230,000 skilled workers by 2050.

Thailand’s ability to capture more value from the AI investment cycle will depend on whether its manufacturing base can support stronger local engineering, domestic suppliers and technology developed within the country.

Source: IEA, BOI, Liquid cooling for AI systems, AI data-centre power and cooling design