EDITORIAL: AI and establishment distrust

Data centers have become one of the fastest-growing infrastructure sectors globally, driven by the rapid expansion of artificial intelligence (AI). However, debates over data centers in the US as well as Asia-Pacific nations have intensified, with public focus shifting from economic growth and technological development to worries about rising costs of living. That has led to construction delays and cancelations.

Opposition to data centers in the US has become a cross-party issue ahead of the upcoming midterm elections. A March Gallup poll found that 71 percent of Americans oppose data centers in their area, with nearly 50 percent saying they strongly opposed them. A Pew Research Center survey in August found that 60 percent of Americans are uncomfortable with new data centers, with 54 percent saying they are concerned about the negative environmental impact, 50 percent about rising utility bills and 49 percent about quality of life.

The US is home to the world’s biggest investors in AI infrastructure and has the most data centers globally. According to Stanford University’s AI Index report, the US has 5,427 data centers, more than 10 times as many as any other country. However, rising opposition has blocked or delayed 45 data center projects in the US in the second quarter of this year, valued at US$68 billion. That followed 75 data center projects valued at about US$130 billion that were blocked or delayed in the first quarter, according Data Center Watch, a research project from 10a Labs that tracks opposition to data center development across the US.

Most of the opposition in the US stems from construction, grid connection and land leasing, which could be resolved over time, Bloomberg reported, citing data compiled by market intelligence platform Currence. Opposition from local communities and politicians is a primary driver of the disruptions, accounting for 52 percent of project cancelations, with regulatory and permission issues accounting for 30 percent, the reports said. Growing public opposition has surpassed financing, grid constraints and critical equipment shortages as the main challenge to data center development.

The growing opposition to data centers, an issue that is uniting Americans across party lines in an otherwise highly partisan era, reflects a more serious concern: public discontent with the political and economic establishments. The Washington Post in July quoted Utah Governor Spencer Cox as saying that data centers embody public anxieties around AI and the economy. Cox — who supports data center development for economic growth and national security, but has called for stricter environmental and consumer safeguards amid public pushback — said data centers have become a tangible target for those anxieties because people feel that “this is something they can fight.”

Analysts from think tanks such as the Brookings Institution have said that while the data-center backlash in the US was triggered by immediate, tangible grievances over potential electricity price hikes, water waste and industrial noise, it was fundamentally fueled by a deeper and broader anxiety about corporate overreach and the unchecked development of AI. Despite the debate over who benefits and bears the costs of data centers, tech giants and US President Donald Trump continue to claim that the expansion is vital for US technological supremacy over China.

Taiwan should watch this issue closely, because grassroots opposition to data centers is spreading to the Asia-Pacific region. Conflicts over data center construction in Japan, Australia and India are sometimes more pronounced than in the US. It remains unclear whether additional restrictions on data centers would weigh on tech giants’ capital expenditure, drive away potential backers and ultimately affect the companies in Taiwan that are in the AI supply chain. Growing policy uncertainties abroad therefore deserve further observation.

The backlash against data centers abroad should not be seen as opposition to AI. Instead, the real question is data center governance, and the answer is that large data centers should bear a larger burden of energy and grid costs because they are big electricity users. Moreover, data center development involves power equipment, cybersecurity, supply chains and national security, but at its core, it is actually about power grid resilience. It is no longer a question of how much electricity can be generated in the AI era, but whether power transmission, substations, energy storage and control systems can withstand new large-scale loads. This is particularly true for Taiwan.