The Democrats' victory in the US midterm elections initially rattled financial markets, with artificial intelligence stocks coming under particular pressure. Yet the subsequent recovery in many technology shares reveals a more complex picture. Investors are increasingly questioning the economics of the AI boom: How long can technology companies continue pouring billions into new data centers when energy requirements, financing costs and the profitability of these investments are coming under growing scrutiny?
The US midterm elections on November 3, 2026, did more than reshape the balance of political power in Washington. The Democrats' unexpectedly decisive victories also triggered significant volatility across international financial markets.
Technology and artificial intelligence stocks were among the hardest hit during the initial sell-off. In Asian trading, Nasdaq 100 futures fell by as much as 2.4 percent, while futures on the broader S&P 500 declined by approximately 1.4 percent at their lowest point.
A recovery followed, however. By early Wednesday morning US Eastern Time, futures on both indices had returned to positive territory. Dow Jones futures also rebounded significantly.
The contrasting performance of individual technology stocks suggests that investors are not abandoning the AI sector altogether. Instead, they are becoming increasingly selective in assessing which companies stand to benefit from technological advances and which face additional economic or regulatory risks.
The Democratic victory has provided a new catalyst for this reassessment. But fundamental doubts about the financing and profitability of the AI boom had already emerged before the election.
The Correction Began Before the Midterms
Even before polling stations opened across the United States, technology stocks had experienced a difficult period.
On November 2, the Nasdaq 100 entered correction territory, having declined by more than ten percent from its record closing high on October 6.
Market observers generally define a technical correction as a decline of at least ten percent from a previous peak. It does not necessarily signal the beginning of a prolonged bear market. It does, however, indicate that investors are increasingly questioning prevailing valuations.
Recent earnings reports from major technology companies had intensified those concerns.
Quarterly results from Alphabet and Amazon renewed the debate surrounding the enormous capital expenditure required for AI infrastructure. Disappointing market reactions to earnings from other technology companies also weighed on sentiment.
Underlying these concerns is a fundamental economic question: Technology companies are investing ever larger sums in data centers, advanced chips and the energy infrastructure needed to operate them. Whether these investments will generate sufficiently high and sustainable returns remains uncertain across several business models.
Growing demand for AI applications does not, by itself, resolve that question.
What ultimately matters is how much customers will pay for AI services over the long term, how operating costs develop and how quickly companies can recover their substantial upfront investments.
Texas Becomes a Political Risk for the AI Industry
The midterm elections have introduced another source of uncertainty.
Particular attention is focused on Texas, where Democrat Gina Hinojosa unexpectedly defeated incumbent Republican Governor Greg Abbott.
Texas has become one of the United States' most important locations for data center development. Large available sites, existing energy infrastructure and a historically business-friendly regulatory environment have made the state particularly attractive to investors.
During her campaign, however, Hinojosa explicitly called for a temporary suspension of new data center approvals until binding safeguards could be established.
Her proposals focus on electricity and water consumption, the impact on local infrastructure and whether communities receive sufficient economic benefits from these investments.
Tax incentives granted to major data center operators are also under scrutiny.
The election result does not mean that previously approved projects will automatically be halted or that a statewide construction ban is already in place. Implementing such measures will depend on existing legal powers, potential decisions by the Texas legislature and the details of individual permitting processes.
Nevertheless, investors now face an additional risk. Projects whose profitability depends on particular tax incentives, low electricity prices or rapid commissioning may have to be reassessed under changing political conditions.
When a Location Advantage Becomes a Cost Problem
The political dispute over data centers has clear economic foundations.
Modern AI systems require enormous computing power. Training large models and processing subsequent requests can involve tens of thousands of high-performance chips.
These chips consume electricity not only for computation itself but also through cooling systems, networking equipment and supporting infrastructure.
According to an assessment by the International Energy Agency, global electricity consumption by data centers could nearly double, from approximately 485 terawatt-hours in 2025 to around 950 terawatt-hours by 2030.
Demand from facilities specifically designed for artificial intelligence is growing particularly rapidly. Their electricity consumption could triple over the same period.
This expansion presents significant challenges for electricity infrastructure. New data centers can be developed faster than additional power plants, substations and transmission lines can be built.
Texas illustrates the scale of the planned expansion. In February 2026, the state's grid operator, ERCOT, reported interconnection requests from large electricity consumers totaling more than 232 gigawatts. Approximately 72 percent of those requests came from data center projects.
These figures represent requested or assessed connection capacity. They are not a forecast that every project will ultimately be completed.
Nevertheless, the substantial gap between proposed facilities and operational sites illustrates the uncertainty surrounding future infrastructure requirements.
When additional transmission lines or generating capacity become necessary, another question arises: Who pays for them?
If the resulting costs increase electricity bills for households and other businesses, an investment project initially welcomed for its economic potential can become a source of local political conflict.
Georgia Changes the Political Landscape Too
Similar debates are taking place in Georgia, where Democrat Keisha Lance Bottoms won the gubernatorial election.
The Atlanta metropolitan area is one of the most important US markets for new data centers. Residents there have also raised concerns about substantial electricity and water consumption and the potential effects on household living costs.
During her campaign, Bottoms likewise advocated a temporary suspension of new data center developments.
The results in Texas and Georgia are particularly significant because they represent more than a potential shift in federal policy.
Although Democrats will control both chambers of Congress and can initiate investigations or propose new legislation, President Donald Trump retains his veto power. At the state and municipal levels, however, authorities can directly influence investment conditions within their respective jurisdictions.
Permitting, water rights, energy infrastructure and tax incentives are regulated partly at the state or local level in the United States.
The AI industry therefore faces more than the prospect of increased scrutiny from Congress. State-level policies may also affect the viability of specific investment projects.
Not All Technology Stocks Are Falling
Wednesday's market movements illustrate why treating the technology sector as a single investment category would be misleading.
While companies directly involved in building or operating large data centers came under pressure in premarket trading, other technology stocks proved considerably more resilient.
Among the weaker performers were IREN, Hut 8, Applied Digital, Nebius and CoreWeave. These companies are involved in various aspects of computing infrastructure and cloud operations.
Performance among the largest technology companies was mixed. Amazon, Microsoft and Meta gained in early trading, while Alphabet temporarily traded lower.
Chipmaker AMD also benefited from an encouraging business outlook, with its shares rising by more than six percent in premarket trading at one point. Nvidia, Micron and Broadcom recorded modest gains during parts of the session.
These price movements reflect early trading conditions, not final daily closing prices.
They nevertheless demonstrate that investors are distinguishing between different business models.
A company financing and constructing data center capacity at its own risk faces different pressures from a chipmaker supplying products to a broad range of customers.
The two business models are still interconnected. If investment in new data centers declines for a prolonged period, demand for high-performance chips could eventually weaken as well.
Conversely, a temporary halt to approvals in certain regions does not necessarily mean that global demand for computing power will decline. Investment could simply shift to other states or countries.
Financing Becomes the Critical Test
One of the central vulnerabilities of the AI boom lies in the timing of its costs and revenues.
Data centers require substantial upfront capital expenditure. Land must be developed, buildings constructed, electricity connections established and powerful servers installed.
Much of this spending occurs long before facilities reach full utilization and generate the corresponding revenues.
Rapid technological development creates an additional challenge. New generations of chips may offer greater computing performance and energy efficiency than existing hardware.
That puts pressure on operators to achieve high utilization rates quickly, before technological advances affect the competitiveness of their existing infrastructure.
As long as demand grows strongly and investors remain willing to provide additional capital, this expansion can continue.
The situation becomes more difficult when several pressures emerge simultaneously: higher financing costs, permitting delays, rising electricity prices and uncertainty about future revenues.
Under those conditions, expectations of long-term growth may no longer be sufficient to justify every new investment.
This is particularly relevant for companies financing part of their infrastructure expansion through debt. If commissioning is delayed, interest expenses may accumulate before the facilities begin generating income.
Investors are therefore increasingly evaluating not only the technical capabilities of new AI systems but also when the billions invested in them will produce meaningful financial returns.
Politics Is Changing Earnings Expectations
The Democratic victory has created a new political landscape. With control of both chambers of Congress, the party can exercise greater oversight of government decisions and introduce additional legislative proposals.
For the technology sector, issues such as data protection, competition policy, artificial intelligence and infrastructure financing could attract greater attention.
However, political intentions must be distinguished from legislation that has actually entered into force.
Donald Trump remains president and can veto new laws. Overriding a presidential veto requires a two-thirds majority in both chambers of Congress.
Comprehensive new federal AI regulation is therefore by no means an automatic consequence of the election.
Financial markets, however, respond not only to laws that have already been passed. Share prices also reflect expectations about future costs, risks and corporate earnings.
The mere possibility of additional requirements or delayed projects can consequently affect the valuation of individual companies.
Distinguishing between actual regulatory changes and shifting perceptions of risk is essential to understanding the current market movements.
Renewable Energy Stocks Benefit from Changing Expectations
While parts of the AI infrastructure industry came under pressure, renewable energy stocks performed considerably better in early trading.
First Solar gained more than eight percent at one point in premarket trading. SolarEdge Technologies and Canadian Solar also recorded increases.
Investors appear to associate the new political balance with the possibility of stronger support for clean energy projects.
Such an outcome is far from guaranteed. Democratic energy legislation could also be blocked by the president's veto.
Moreover, political opposition to new data centers does not necessarily imply weaker demand for renewable energy.
On the contrary, if electricity supply becomes a decisive constraint on AI expansion, additional power generation, energy storage and grid infrastructure could become increasingly important over the longer term.
The financial performance of companies operating in these sectors will nevertheless depend on actual orders, investment costs, regulatory conditions and their own competitiveness.
The AI Boom Faces a New Test
The US midterm elections have made the political risks facing the technology sector more visible. They did not, however, create the fundamental economic challenges that the industry is now confronting.
Even before election day, high capital expenditure, disappointing earnings expectations and the growing use of debt to finance infrastructure had raised doubts about certain company valuations.
Political resistance to new data centers now adds another layer of uncertainty.
At the same time, the rapid recovery in numerous technology stocks demonstrates that investors are not anticipating a universal end to AI growth. Demand for computing power remains strong, and many companies continue to expect substantial economic opportunities.
The increasingly important question is which providers can operate their investments profitably under real-world conditions.
For investors, the focus is shifting. For years, attention centered on how quickly new data centers could be constructed and how many advanced chips could be sold.
Looking ahead, electricity costs, grid access, permitting risks, financing arrangements and actual utilization rates are likely to play a greater role in determining financial success.
The midterms have accelerated this change in expectations. Whether it results in a prolonged revaluation of the technology sector will depend primarily on future corporate earnings and the actual implementation of political measures.
The AI boom does not necessarily have to end because its operating environment is changing. But for parts of the industry, the days when ever-larger investment announcements alone were enough to justify rising valuations may be becoming more difficult.
SK
