The US Democratic Party swept both chambers of Congress in the midterm elections, fueling market concerns about the outlook for US data center construction. But Nomura Securities argues that the net impact of this political shift on China's AI economy may actually turn out to be positive.
According to a report published on November 6 by Nomura economists Ting Lu and Jing Wang, the Democratic "blue wave" will slow US data center expansion through tighter regulation and electricity price protections, which would in turn trigger a pullback in global AI chip prices and ease the pressure on Chinese companies in securing advanced chips. "We expect China's AI economy as a whole to benefit significantly from the US blue wave," the report stated.
Market reaction already tells the story. Within two trading days after the election results came out, the Nasdaq Composite fell 5.5% in a single day on the second trading session, marking its largest one-day drop in 18 months. The Philadelphia Semiconductor Index (SOX) tumbled nearly 15% over the two days. NVIDIA dropped about 12% in total over the two sessions, Micron Technology fell roughly 24%, while SK Hynix and Samsung Electronics declined 9.5% and 7.5% respectively on South Korea's KOSPI market, triggering automatic circuit breakers.
Morgan Stanley, in a report released around the same time, noted that Democratic control of Congress does not mean a substantive easing of US-China relations, but rather an improvement in policy predictability. On AI investment strategy, the firm favors China's "self-reliance and controllability" theme over export-chain names that depend on global AI capital expenditure.
US Data Center Construction: Tighter Regulation, Expansion Hindered
The most direct impact of the election results is that US data center construction faces stricter policy constraints.
According to the Nomura report, Democrats won governorships in several key states for data center construction, with the new leadership heavily emphasizing consumer electricity cost burdens, energy affordability and corporate accountability during their campaigns. This means data center developers will no longer be able to pass on the capital costs of grid upgrades, transmission expansion and new power generation to residential users—the financial burden will fall directly on the developers themselves.
The firm pointed out that stricter municipal oversight, bans on signing confidentiality agreements during the early site selection stage, and mandatory clean energy procurement requirements will replace the previous rapid industrial development model. "The time required for a new hyperscale data center from site selection to grid connection will be significantly extended," the report said.
Currently, the scale of US data center construction is enormous. According to tracking data from Cleanview, there are more than 4,500 data centers across over 1,500 campuses in the US, of which about 1,418 are already operational, 564 are under construction, and another 2,555 are in the planning stage. Data from commercial real estate firm Cushman & Wakefield shows that capacity currently under construction has reached a record 37.7 gigawatts, with the planned pipeline reaching as high as 285 gigawatts.
Chip Price Decline: An Unexpected Dividend for China's Data Centers?
For China, the most direct benefit of a slowdown in US data center construction is reduced difficulty in obtaining AI chips and eased price pressure.
The firm noted that "when US cloud giants aggressively expanded data centers, they hoarded global capacity for GPUs, HBM and advanced logic chips, tightening global supply and driving up premiums for restricted hardware in black markets and secondary channels."
Once US data center capital expenditure slows, this situation will ease. The firm believes Chinese AI companies, cloud service providers and research institutions will be able to obtain relevant GPUs, HBM modules and advanced packaging materials more easily and at lower cost through intermediary trade hubs.
The compression of gray market premiums is particularly noteworthy. Analysts pointed out that during the peak of US data center construction, gray market premiums for advanced GPUs and HBM modules at Asian logistics hubs often reached 200% to 300% of list prices. As Western demand cools, "these speculative premiums could collapse significantly, enabling Chinese buyers to acquire computing hardware at more reasonable prices."
At the infrastructure level, China is advancing a state-led national AI data center network. The 15th Five-Year Plan includes plans to invest approximately 2 trillion yuan (about $295 billion), with the goal of integrating scattered regional computing facilities into a unified, interconnected network by 2028, operated primarily by state-owned telecom giants such as China Mobile and China Telecom. Unlike the local resistance facing US grid construction, China relies on the "East Data, West Computing" project, coordinated uniformly by the State Grid, allowing western computing hubs in Guizhou, Inner Mongolia, Gansu and Ningxia to directly access low-cost renewable energy without dealing with residential electricity price disputes or grid connection queues.
Morgan Stanley: Bullish on China's Localization and "Self-Sufficiency" Theme
Morgan Stanley provided a supplementary perspective on the above logic in its report released around the same time, though with a more cautious tone.
Strategists Laura Wang and Vicky Wu pointed out that Democratic control of Congress should not be interpreted as a substantive easing of US-China relations. "We believe strategic competition and geopolitical uncertainty will persist."
The strategists believe the potential positive impact of this election lies in the improved predictability brought by policy constraints. Congressional checks and balances will make large-scale unilateral policy shifts more difficult, reducing the probability of escalating tariffs becoming a primary bilateral policy tool.
The firm suggested distinguishing between two types of China AI trades: beta exposure to the global AI supply chain (such as optical transceivers, which mainly follow global AI capital expenditure and US policy scrutiny), and China's localization and self-sufficiency theme.
"We prefer the self-sufficiency trade—it enjoys both supply and demand support within the framework of China's national policy priorities." Morgan Stanley believes that continued bipartisan technological competition with China will only strengthen China's domestic policy commitment to technological self-reliance, covering semiconductors, computing power, power equipment and domestic AI infrastructure.
Additionally, Morgan Stanley noted that compared to the first half of 2026 when the market was overly concentrated in AI trades, investment opportunities in China's stock market are now more balanced, covering consumer and industrial exporters, oil price-sensitive downstream sectors, healthcare/biotech, and green energy sectors. The report also noted that China's stock market is currently valued near the historically absolute low range of about 10%, and various funds have recently significantly increased their buying of A-shares.
Comments