Post-Election US Data Center Policy Adds Uncertainty
The US midterm elections have concluded, with Democrats winning in both chambers of Congress and key gubernatorial races, bringing policy uncertainty to expanding artificial intelligence data centers. Local backlash over data center electricity and water use has intensified, and markets worry momentum will be curbed for emerging cloud service providers and chip hardware makers such as NVIDIA, while large cloud giants such as Microsoft, Amazon and Alphabet have instead gained breathing room. The concerns, combined with a US stock pullback, directly hit Taiwan stocks reliant on AI servers, with panic selling on the morning of the 6th, an intraday drop of up to 3737.67 points, the largest intraday drop on record, and benchmark leader TSMC plunging at one point to the limit-down price of 2020 NT dollars, while Hon Hai, Quanta and other frontline server contract manufacturers fell sharply in tandem, and the weighted index closed down 1873.08 points at 39701.64 points, losing the 40000-point mark.
Local Calls for Moratorium and Analyst Views
A column in tech media The Information noted that markets had originally expected Democratic control of Congress to help cool geopolitical conflicts and reduce tariff uncertainty, but the election results instead brought many officials skeptical of overheated AI development and data centers into office. In key regions such as Texas, Florida, Ohio and New Jersey, multiple local referendums and candidates have called for moratoriums on building data centers. The report cited senior Cantor Fitzgerald analyst Muse in an interview on TITV: "I hope calm and rational people will eventually realize how important AI computing power is." He believes that once local politicians recognize the substantial contribution of data centers to local jobs and tax structures, opposition will gradually shift toward pragmatic negotiation.
Giants and Emerging Cloud Diverge
Concerns over an infrastructure slowdown have split the tech camp. Analysts noted that CoreWeave, Nebius and other emerging cloud providers focused on graphics processor clusters, lagging in self-built facility progress and reliant on rapid compute expansion, will bear the brunt in revenue momentum, while semiconductor hardware makers such as NVIDIA, Micron and SK Hynix also face valuation corrections. In contrast, hyperscale cloud operators such as Microsoft, Amazon and Google parent Alphabet, with large existing facilities forming a moat, have gained a reason to slow hefty capital spending, prompting funds to flow back against the trend. For the Taiwan supply chain, hyperscalers deploy both NVIDIA GPUs and self-developed application-specific chip architectures, relying on full-rack delivery and global-scale manufacturing capabilities, making TSMC, Hon Hai, Quanta and Wiwynn core pillars, while emerging cloud operators focused on GPU compute rental rely on short lead times and flexible customization, opening entry points for server and board makers such as Gigabyte, ASUS and ASRock.
Tung Tzu-hsien Says Worries Are Overblown
On concerns that the Democratic win and local opposition could trigger a slowdown in construction, Pegatron Chairman Tung Tzu-hsien said worries were "overblown." He said AI had already fully risen late in the Biden administration, is not the product of a single party, but is central to sustaining US global hegemony and technological competitiveness, and no government will let AI development stall. He stressed that Taiwan's tech industry is highly irreplaceable in the global supply chain and should not easily change course because of US election results or tariff threats; the most important thing is to focus on core business and move forward steadily. Facing expanded post-election foreign-investor selling of Taiwan stocks, he described foreign capital as "nomads" chasing fertile pasture, saying short-term inflows, outflows and hedging do not mean fundamentals have weakened, Taiwan has always had "abundant pasture" with sound fundamentals, and funds will eventually return after settling.