The US midterm elections, in which the Republicans suffered a heavy defeat, managed what few thought was possible. They humbled President Donald Trump, even if only momentarily. The artificial intelligence industry, which had seemed oblivious to all questions about its neck-breaking speed, suddenly found itself checked. Lastly, for investors and businesses in India, there was a two-way sense of relief.
First, it no longer looked so bad that the country did not have an AI story to sell to investors, even as emerging market peers such as South Korea and Taiwan rode the boom. Second, India’s AI infrastructure—especially its data centres—may see an even heavier influx of foreign investment. The larger impact could be even more lasting and profound.
It was no coincidence that a vast majority of AI infrastructure was concentrated in Republican-led states. About 75 per cent of all AI data centres under construction and over 80 per cent of those planned were in Texas, Georgia and Ohio. Texas was said to be the ‘epicentre’ of this AI boom and was set to host a quarter of all planned AI infrastructure in the US. That outlook no longer holds.
The markets sensed this shift. Following the results, shares of AI and semiconductor companies tested rock-bottom levels. Shares of AMD fell 17.3 per cent in a single day. Micron fared even worse, dropping 17.5 per cent, while Nvidia lost $450 billion in market value in just one trading session. This occurred even before details surfaced regarding its founder Jensen Huang’s search for sex outside his marriage and that it could have ended in an alleged honeytrap planned by the Chinese.
More questions are bound to follow. “The sharp pullback in global markets was driven by AI momentum plays getting sold into concerns over valuation, overinvestment, and a lack of credible visibility on returns from massive AI investments,” says market expert Ajay Bagga.
These are critical issues. As regulators and investors in the US look for greater clarity, the Indian market will increasingly present a natural hedge against the AI bet.
Indian markets could undergo a structural shift in the coming months. “With institutional allocations to India recovering from underweight levels, global money is rotating into India’s structural growth story. Brokerages estimate that any catch-up reallocation to underweight Indian markets could see huge FPI inflows crowding into the Indian market,” says Bagga.
Heightened scrutiny
The US remains the central player in the global AI stack. It is home to the world’s leading AI foundation models, built by the likes of OpenAI and Anthropic, as well as major hardware players like Nvidia and AMD. The US is also where much of the world’s computing power resides.
The standards and regulations set by the US for these companies default to becoming global norms. Supply chains, infrastructure, export policy, and data security laws are all influenced by developments in the American legislature and courts. Then, of course, there is the financial aspect: roughly three-quarters of global investment in AI flows into the US.
It would not be entirely wrong to say that AI players enjoyed a free hand until now, given the close relationships leading AI entrepreneurs cultivated with the Republicans and Trump. They also attempted to influence the midterm elections. Elon Musk, CEO of electric carmaker Tesla, poured $500 million into the campaigns. Almost half of that went to Texas, specifically supporting Republican candidate Ken Paxton in his Senate bid. However, those dollars could not buy votes, and Paxton lost to Democrat James Talarico.
This is not to say there has been no regulatory action at all. Although Congress has not passed a single defining federal AI law, state governments and authorities have introduced several regulations. These have focused on digital forgeries, defence and federal security, anti-deception laws targeting AI vendors, and specific health care interventions. In many ways, the states compensated for the federal government’s light-touch approach.
Why, then, was public anger against AI so intense that it became a major issue in the midterm elections? “As has been the case with every technology, AI will gradually become cheaper and more accessible. Even with tighter regulation, the cost of using AI technology will fall. But that does not mean the overall ‘cost’ to society will decrease,” says Russell A. Stamets, an American lawyer practising in India.
The ‘costs’ Stamets refers to include mounting public concern over high water consumption, soaring energy bills, fake or manipulative news, and job security—issues that have now become household concerns across the US. Voters felt that none of these had been adequately addressed by existing laws and regulations.
“Democrats are known for regulation, and it is no secret that they will introduce more restrictions,” says Peter Cardillo, chief market economist at New York-based Spartan Capital Securities. “But will they do away with AI or data centres? I don’t think so. Trump is still the president and holds veto power. He will attempt to work with the Democrats to limit any regulation that could severely impair the AI ecosystem.”
Trump will undoubtedly play hardball, but the ground reality has fundamentally shifted. Stamets, a partner at the law firm Circle of Counsels, suggests that regulation will now focus heavily on the environmental impact of AI, alongside data security and monitoring concerns. This shift, he argues, will also ripple through to India.
The core issue
At the core of the popular vote against AI lies a fundamental question that every wave of industrialisation has had to answer: who pays the price for progress?
In the US, local communities have seen their electricity bills rise to offset the pressure placed on power grids by the heavy demands of data centres. While the average price increase ranges between 2 per cent and 6 per cent, it has spiked to 14 per cent in some states. Estimates suggest energy costs could rise by up to 30 per cent by 2028.
The second issue is water. A single data centre can consume roughly half a million gallons—or 1.8 million litres—of water per day. That is enough to meet the daily drinking water needs of 7.5 lakh people. In the US, this has become a flashpoint because 20 per cent of data centres are situated in regions experiencing moderate to severe water scarcity.
Given the severity of these issues, Stamets believes environmental impact will be a primary focus of upcoming AI regulations. “Tech companies will be required to provide greater transparency regarding their resource consumption. Democrats are expected to push for stricter limits on new data centres, and some states may revoke tax exemptions for them. It will be a gradual process, but data centres will have to become cleaner,” he explains.
The second major pillar of AI regulation will be security, which encompasses multiple layers. For instance, new iterations of models like Anthropic’s Claude or Google’s Gemini may require mandatory third-party safety audits. Regulations will also zero in on consumer protection, including safeguards against the algorithmic exploitation of children. “Tech companies should expect to be called in for individual inquiries and legislative investigations,” Stamets adds.
Another feature will be the ongoing friction between state and federal authorities. ‘Unlike India, the US does not have an explicit Central List, State List, or Concurrent List. But to borrow that framework, one could say Trump had placed AI on the central list. That is now set to change; it will move, at the very least, onto the concurrent list,” says Stamets.
However, Democrats will need to handle their newly won leverage carefully and strike a balance. “AI accounts for about 2 per cent of US GDP, so you would not want to hit it too hard,” says Cardillo. Furthermore, research by J.P. Morgan Global Research and Asset Management indicates that the broader AI ecosystem—including investments in data centres and tech hardware—has driven nearly one-fifth of US economic activity over the past year. Democrats will have to ensure that their system of checks and balances does not stifle growth.
Impact on India
India remains a much smaller market for data centres by comparison. Texas alone houses more data centres than the whole of India. However, Indian numbers are expanding rapidly. The country currently hosts around 350 data centres with a combined capacity of 1.5 GW, a figure expected to reach 8 GW by 2030, backed by investment commitments of nearly $100 billion. Maharashtra leads the domestic market with roughly a 30 per cent share, followed by Tamil Nadu, Karnataka, and Telangana.
“Data centre capacity in India is expanding rapidly due to strict data localisation regulations, massive 5G rollout consumption, and surging demand for localised AI workloads,” says Bagga. He adds that India stands to benefit from redirected investment flows should data centre rollouts slow down in the US.
A broader impact will be felt across the entire AI stack. “AI sentiment is undergoing a fundamental shift—from speculative hype to a focus on ROI and execution. Early-stage AI hardware providers and chip capital supply chains face near-term valuation pressure as hyperscaler capital expenditure peaks. However, AI-enabling infrastructure, data engineering firms, and enterprise software implementation providers remain long-term beneficiaries as actual deployment replaces hype,” Bagga explains. Energy suppliers—ranging from power plant equipment manufacturers and power generators to distribution companies—are also set to benefit from the sustained demand generated by AI data centres.
Beyond the tech sector, Indian businesses could benefit from a divided or policy-neutral US Congress. “This will help stabilise international trade tariffs and keep corporate taxation rates predictable. Indian IT services, generic pharmaceutical exporters and electronics manufacturers stand to benefit directly from a stable dollar-rupee dynamic and reduced cross-border regulatory friction,” says Bagga.
Despite these potential gains, Indian policymakers and industry leaders must work to ensure that AI does not become a contentious political issue domestically. “As awareness grows among the Indian public, I expect they will ask the same fundamental question: is AI truly good for us?” says Stamets. Protests have already surfaced in Thane, Maharashtra—the proposed site for a major Amazon data centre—where local concerns closely mirror those in the US: fears over excessive water and power consumption straining community supplies.
The ideal outcome for India would be to draw lessons from the US experience, putting early guardrails in place to protect local communities and consumer rights while striking a balance with the growing demand for AI capacity.
