Shares of global semiconductor and data center infrastructure companies fell sharply as concerns spread that regulations on AI data center construction could tighten after the U.S. midterm elections.
In South Korea’s stock market on November 5, Samsung Electronics closed at 201,500 won, down 7.57% from the previous trading day, while SK Hynix ended at 1,264,000 won, down 9.46%. The Kospi also plunged 5.28% to close at 5,383.80.
This mirrored a divergence among sectors in the U.S. stock market the previous day.
On November 4 local time, the S&P 500 rose 1.58% and the Nasdaq Composite gained 1.44%, but the Philadelphia Semiconductor Index fell 3.13%. Among major stocks, Nvidia (-3.17%) and Micron (-7.00%) declined, while Microsoft (+6.68%) and Amazon (+6.12%) posted strong gains.
The simultaneous rise in major indexes and decline in semiconductor stocks suggests investors have begun assessing not just the AI industry’s overall growth potential, but also the pace of data center expansion and the investment burden on individual companies, by different standards.
Prospect of tighter regulation grows in major data center hubs
In the U.S. midterm elections, Democrats secured majorities in both the Senate and the House of Representatives. Democratic candidates also won gubernatorial races in Texas, Ohio and Georgia, where data center development is brisk.
In particular, Texas governor-elect Gina Hinojosa said during the campaign that she would halt approvals for new data centers until laws protecting local residents were in place, and scale back tax breaks for data center operators.
Ohio governor-elect Amy Acton called for development to proceed only if it met requirements on utility costs and labor and environmental standards. Georgia governor-elect Keisha Lance Bottoms pledged a temporary halt to development to assess the impact of data center construction on local communities.
These states are key markets where demand for AI data centers is growing rapidly. Restrictions on new permits or connections to the power grid, or rising development costs, could affect related investment schedules.
According to a Goldman Sachs analysis cited by Reuters, areas where Democratic victories in state elections have raised the prospect of tighter regulation account for about 45% of planned U.S. data center capacity, which could face temporary construction halts or regulatory changes.
The impact of any regulations could also vary depending on federal and state government powers, the legislative processes in each state and the validity of permits already issued.
Underlying these developments is growing opposition from local communities over the electricity, water resources and public infrastructure costs required to build data centers.
In a survey released by polling firm Ipsos in September 2026, 69% of U.S. adults expressed concern about the expansion of AI data center construction. Among those who were concerned, the main issues cited were the strain on local water supplies (50%), the strain on power grids (41%) and the possibility of higher household utility bills (35%).
Ultimately, the debate over data center development has expanded beyond whether to foster the AI industry to how much of the cost of expanding infrastructure for large facilities should be borne by companies and how much by local residents.
Conflicting interests of semiconductor suppliers and Big Tech
If construction of new AI data centers is delayed, the companies most directly affected may be those supplying the equipment and infrastructure they require.
Building data centers requires semiconductors, including GPUs and HBM, as well as servers, networking equipment, power supply equipment and cooling systems. Project delays could also affect orders and delivery schedules for these products.
Neocloud companies, in particular, build new facilities or sell secured computing capacity as cloud services. Their results can vary considerably depending on capital expenditure, financing and customer contracts, making them especially sensitive to changes in investment schedules.
For hyperscalers such as Microsoft and Amazon, which already have large data centers and established cloud customer bases, the calculation may be different.
These companies have spent heavily on capital expenditure (CAPEX) to expand their AI infrastructure. If the pace of investment slows, reduced cash outflows for building new facilities could have a positive effect on free cash flow (FCF).
If the supply of new data centers is also constrained, the relative value of operating facilities that have already secured power and permits could rise—another factor seen as favoring established large operators.
Of course, for reduced investment to deliver a real improvement in cash flow, existing facilities must be able to meet customer demand even if data center construction is delayed. Conversely, if a shortage of computing resources limits AI service supply and revenue growth, Big Tech would also find it difficult to avoid a negative impact.
How might the growth outlook change for South Korean semiconductor companies?
The decline in semiconductor stocks that began in the United States spread to the South Korean market. On November 5, foreign investors net sold approximately 3.4248 trillion won on the main stock market, adding to downward pressure on the KOSPI.
As Samsung Electronics and SK hynix are key companies in the supply chain for AI data center memory chips, changes in capital expenditure by global cloud operators could be an important factor for them.
Demand for high-bandwidth memory (HBM) and high-capacity server memory used in AI servers is tied not only to data center construction plans but also to GPU production, equipment additions at existing facilities and demand for running AI models.
Key factors shaping the market’s direction will be how far the possibility of tighter regulation translates into actual policy, whether delays in data center development amount to no more than schedule adjustments, and how much any reduction in Big Tech’s capital expenditure is reflected in semiconductor suppliers’ results.
