Beware the political cost of data centres

 The push for data centres has wrought huge environmental costs and increased pressure on energy resources. This can be a political minefield, as seen in the US.

The global race to build artificial intelligence (AI) data centres is entering a politically sensitive phase. Those in the United States just witnessed a massive electoral recoil. The Republican party was thrashed. Badly.

Hence, Malaysia must understand the implications before becoming an alternative destination for facilities facing growing resistance in the United States.

As they head to Malaysia, Putrajaya must understand these data centres come with huge political costs as President Donald Trump has discovered on Nov 3, 2026 in the mid-term election. Why?

For much of Trump’s second administration, the White House has encouraged American technology giants to accelerate investments in AI infrastructure.

The underlying assumption was straightforward.

Whoever controls the largest computing capacity will enjoy a decisive advantage in the development of advanced AI, including potentially super intelligent systems.

Yet technological ambition does not automatically translate into public acceptance.

Across the United States, opposition to hyperscale data centres has grown over electricity consumption, water requirements, noise pollution, land use and the possibility of higher utility bills.

These concerns have become politically significant.

Democratic politicians, including Gina Hinojosa in Texas, Amy Acton in Ohio and Keisha Lance Bottoms in Georgia, have called for pauses on major data centre developments.

Democrats have also discussed congressional investigations into technology companies, raising the prospect of greater scrutiny over the environmental and economic consequences of AI expansion.

The significance extends beyond American domestic politics.

When the world’s largest AI market encounters resistance to physical infrastructure, international investors will reassess where future computing capacity can be built.

Malaysia is an obvious candidate for consideration. International trade and industry minister Johari Ghani is well aware of the risks based on his public statements though.

Malaysia’s geographical position, proximity to Singapore, expanding digital infrastructure and relative political stability make it attractive to multinational technology companies.

Johor has already emerged as a major data centre destination, while investments in Selangor and other states are expanding Malaysia’s digital footprint.

But Malaysia must distinguish between attracting high-quality digital investment and becoming a convenient location for infrastructure that other countries increasingly regard as environmentally burdensome.

A hyperscale data centre is not an ordinary manufacturing facility.

It can consume enormous amounts of electricity continuously, require substantial cooling infrastructure and place additional demands on water resources.

The economic benefits are also uneven.

Construction creates employment, but permanent staffing requirements can be relatively modest compared with the scale of capital invested.

Moreover, expensive imported semiconductor chips and specialised equipment may account for a substantial proportion of investment expenditure.

Malaysia therefore cannot assume that every billion-ringgit investment announced translates into an equivalent expansion of domestic technological capabilities.

The government must ask whether these investments strengthen Malaysia’s AI sovereignty or merely turn the country into a host for foreign-owned computing infrastructure.

This distinction is crucial.

If Malaysian electricity, water and land support AI systems whose intellectual property, strategic decisions and profits remain overwhelmingly overseas, the long-term national benefits may be limited.

Conversely, carefully negotiated investments could strengthen local research, engineering capabilities, cloud services, cybersecurity and the development of Malaysian AI models.

The difference lies in the conditions imposed before approval. Malaysia should require credible assessments of electricity demand, water consumption, grid capacity and environmental sustainability.

It should also establish transparent mechanisms to ensure that ordinary households and small businesses do not indirectly subsidise infrastructure used by exceptionally wealthy technology corporations.

Renewable energy commitments must be verifiable rather than promotional.

Equally important, data centre investments should generate meaningful technology transfers, skilled employment and opportunities for Malaysian universities and enterprises.

Malaysia must avoid repeating the familiar pattern of welcoming foreign capital without sufficiently strengthening domestic technological ownership.

There is also an Asean dimension.

Singapore, Indonesia, Thailand and Vietnam are competing to attract digital investments.

Competition without coordination could encourage governments to offer excessive incentives while overlooking environmental costs.

Asean should instead develop common principles governing sustainable AI infrastructure.

These could include energy efficiency, water security, cybersecurity, data protection and responsible management of increasingly autonomous AI systems.

Malaysia can contribute to such regional discussions while maintaining an investment environment that remains competitive.

The American backlash against hyperscale data centres is not necessarily a rejection of artificial intelligence itself.

It reflects a growing distinction between the benefits promised by AI and the physical costs imposed on communities hosting its infrastructure.

Malaysia should learn from that distinction.

The country needs foreign investment, but it also needs affordable electricity, reliable water supplies, environmental protection and technological independence.

These objectives are not mutually exclusive. They require disciplined planning and regulatory enforcement.

Malaysia should welcome AI investments that deepen its technological capabilities, not simply those seeking cheaper land, electricity and regulatory approval.

The United States is discovering that the infrastructure underpinning AI carries political and social costs.

Malaysia should not wait until those costs become domestic controversies before responding.

The next generation of AI data centres may indeed look increasingly towards Southeast Asia.

Malaysia must ensure that becoming a preferred destination does not mean inheriting problems that others have begun to resist.

Otherwise, the federal and state governments both risk a voters’ revolt as seen in the United States. - FMT

The views expressed are those of the writer and do not necessarily reflect those of MMKtT