What explains this stark divergence—and what does the US experience forewarn for Asia as its own infrastructure pipeline hits a record 26.5 GW ?
The core difference is not that data centres in the US are physically more harmful. It is institutional and political.
The US has strong local land-use zoning authority, an active civil-society protest culture, and low public trust in large tech companies—all of which turn resource anxieties into organised political action. In Asia, many governments centralise land-use decisions, limit public consultation, position data centres as national competitiveness projects, and benefit from weaker independent media and civil society . However, as the Malaysian case shows, once resource strains become visible, discontent can surface quickly even in less activist contexts.
The grievance list is concrete and backed by polling.
Water consumption is the top-cited fear . Data centres use enormous volumes of water for cooling. In Johor, Malaysia, 47 data centres were found to consume approximately 675 million litres of water daily—enough to fill 270 Olympic swimming pools .
Air and noise pollution follow closely. Diesel backup generators, 24/7 cooling fans, and construction dust degrade nearby quality of life. In Washtenaw County, Michigan, residents filed a class-action lawsuit over noise exceeding 50–62 dB, well above WHO nighttime guidelines .
Electricity price hikes and grid strain are major fears—that massive power draw will raise residential rates and cause blackouts . Farmland loss is another flashpoint, as facilities are sited on agricultural or rural land with little local say
.
Notable flashpoints in 2026:
Political consequences have been dramatic. Opposition has become one of the few unifying issues in a deeply polarised country . Voters have launched recall campaigns against local officials who approve projects
. At least 75 data centre projects worth over $130 billion were delayed or cancelled in the first three months of 2026 alone
. Active opposition groups ballooned from 396 at the end of 2025 to 833 by March 2026, spanning 49 states
.
Strikingly, only 8% of US opponents actually live near a data centre, according to a Milltown Partners survey . The opposition has gone national.
Asia's APAC data-centre pipeline hit a record 26.5 GW in H1 2026, with Malaysia overtaking Singapore as the region's fastest-growing hub . The pipeline now comprises 4.8 GW under construction and 21.7 GW in planning . But the resource strain is becoming visible.
The February 2026 protest in Iskandar Puteri, Johor, involved 50 residents who claimed to represent 1,000 households from four housing estates . Their concerns: dust pollution from construction, health risks, and fears that the facility would impact their water supply . The project—a 300-megawatt complex—was being built less than a kilometre from homes .
In response, the Johor state government took action. It classified Tier 1 and Tier 2 data centres as high water users, halted new approvals for that category, and began requiring new projects to disclose their electricity source . In Selangor, Malaysia's richest state, debate about the burden on communities has become politicised
.
That is the same dynamic that scaled up in the US.
As Asia's data centre buildout accelerates, the US experience offers five clear lessons.
Lesson 1: Resource transparency cannot wait. US backlash accelerated because communities learned about water and electricity consumption after projects were approved. Asian governments should mandate public disclosure of water use, power demand, and emissions before permitting.
Lesson 2: Early, meaningful community engagement is cheaper than crisis management. In the US, opposition shifted from local zoning fights to national protests within two years . Asian governments should establish formal community consultation mechanisms. Ignoring local consent leads to project delays and cancellations worth billions
.
Lesson 3: Treat data centres as public infrastructure, not private real estate. Analysts argue Southeast Asia should plan data centres at a system level—integrated with national power grids, water systems, and land-use plans—rather than project by project . The US patchwork of county-level approvals created inconsistency and conflict.
Lesson 4: Pre-empt the NIMBY tipping point. Only 8% of US opponents actually live near a data centre, yet opposition is national . In Asia, the gap between resource anxiety and protest may be narrow. Malaysia's February 2026 protest—50 residents claiming to represent 1,000 households—shows the same dynamic that scaled up in the US
.
Lesson 5: National competitiveness arguments are wearing thin. US governors initially sold data centres as economic development but now face recall elections . Asian governments should not rely solely on the "AI race" narrative to justify projects. Local populations will expect tangible benefits and measurable environmental protections.
The US data centre backlash did not emerge from nothing. It was fuelled by a lack of transparency, weak local consent, and a perception that the costs of AI infrastructure were being borne by communities while the benefits flowed to distant corporations. Asia is not immune to those dynamics. Malaysia's first protest in 2026 is almost certainly not the last.