Could U.S. Power and Permit Hurdles Disrupt Korea's HBM Shipments?
The November 2026 U.S. midterm elections have highlighted growing local opposition to artificial intelligence (AI) data center development. Concerns over electricity prices, water consumption and environmental impacts are influencing local permitting decisions, creating new uncertainty for Big Tech's infrastructure expansion. Construction delays could affect when servers and AI accelerators are…
The U.S. midterm elections in November 2026 have sparked local opposition to artificial intelligence (AI) data center projects, with concerns over electricity costs, water usage and environmental effects influencing permitting decisions. This uncertainty is impacting the expansion plans of major technology companies, potentially delaying the deployment of servers and AI accelerators that would drive future high-bandwidth memory (HBM) demand for Samsung Electronics and SK hynix.
The core issue revolves around the discrepancy between announced AI investments and the necessary grid connections and regulatory approvals to operate data centers.
According to Ballotpedia's November 6 analysis, voters supported restrictions on data center development in 40 of 41 local ballot measures on November 3. Of those, 27 measures banned or opposed development, while 13 measures rejected proposals favorable to construction. Some communities implemented voter approval requirements for projects using significant land, electricity or water resources.
The elections also saw political shifts in key data center markets, with Democratic candidates winning gubernatorial races in Texas, Ohio and Georgia. Texas Governor-elect Gina Hinojosa has pledged to delay new data center approvals until legislation for additional community protections is enacted.
While these local changes do not constitute a nationwide halt to data center construction, they underscore the growing political and regulatory risks for new projects, especially in regions concerned about infrastructure expenses and environmental impacts. Reliable electricity supply is as critical to data center development as servers and semiconductors, and AI data centers demand substantial power delivery due to their high-performance GPUs and cooling systems.
The International Energy Agency (IEA) projected that global data center electricity consumption would rise from 415 terawatt-hours (TWh) in 2024 to 945 TWh by 2030. The IEA also identified grid access, equipment availability and permitting delays as significant constraints on data center expansion, estimating that building new transmission lines in advanced economies could take four to eight years and that about 20% of planned data center projects might face delays if grid constraints are not addressed.
Tighter development restrictions may impact AI infrastructure companies differently. Those with established data centers, secured power agreements and approved expansion projects might have more flexibility than firms relying on new construction and external financing. Companies such as Microsoft, Amazon and Alphabet could potentially increase utilization at existing facilities or shift server deployments to other locations.
However, their flexibility is constrained by available electrical capacity, cooling systems and network connectivity. Conversely, companies pursuing rapid expansion, including Oracle and CoreWeave, should closely monitor construction schedules and financing conditions, as delays could extend the payback period for infrastructure investments by increasing financing costs and postponing revenue-generating operations.
For Korean HBM suppliers, the timing of new orders may be influenced by these U.S. data center restrictions, as construction delays do not automatically translate into immediate HBM order cancellations or reduced semiconductor revenues.
Written by urgent.news from Korea IT Times's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.