A New Kind of Electricity Shock
In 2025, American households faced an unexpected shock. Electricity bills were climbing by double digits, even as families consumed less power. In Maine, usage fell by around 7 percent compared to the previous year, and in New Jersey, by 6 percent. Yet in both states, the average household electricity bill rose by nearly 17 percent.
This paradox — paying more for less — puzzled many. But experts traced it to an invisible mechanism in the U.S. electricity market: capacity auctions. These auctions, designed to guarantee grid reliability, have become the vehicle through which households are indirectly funding the energy hunger of Big Tech’s sprawling artificial intelligence (AI) data centers. This phenomenon has been aptly described as a “Hidden AI Tax.”
The U.S. System: Capacity Auctions Explained
To understand how this hidden tax operates, one must first grasp the difference between two markets in the U.S. power sector.
Energy Market: This is straightforward. Power plants sell electricity by the kilowatt-hour (kWh), and customers pay for what they actually use.
Capacity Market: Here, plants are paid not for actual electricity delivered but simply for being available if needed. In other words, they are compensated for “standing by.” Even if a household uses less power, these costs are still collected and passed on in their bills.
Grid operators such as PJM Interconnection (which covers New Jersey and 12 other states) and ISO-New England (which covers Maine and neighboring states) conduct annual capacity auctions to ensure that enough reliable supply is always on call.
For years, these auction prices were modest. In PJM, for example, clearing prices stood at just $28.92 per megawatt-day in 2024. But in the very next auction for the 2025–26 year, prices shot up nearly tenfold to $269.92/MW-day, and then jumped again to $329.17/MW-day for 2026–27. In New England, Forward Capacity Auctions cleared at around $2.61 per kilowatt-month — a 31 percent increase over the prior round.
The result: capacity costs ballooned, and households were saddled with higher bills, regardless of their actual consumption.
The AI Boom and Data Centers’ Hunger for Power
What triggered such extraordinary spikes? The most important factor is the explosion of AI and hyperscale data centers.
Training advanced AI models like GPT-4 or GPT-5 requires thousands of graphics processing units (GPUs) operating around the clock. Each GPU consumes as much electricity as a high-end appliance. Multiply this by tens of thousands, and a single AI data center can demand 100 to 500 megawatts — equivalent to the load of a small city.
In Northern Virginia, known as “Data Center Alley,” Amazon, Microsoft, Google, and Meta operate vast server farms. Analysts note that data centers account for the overwhelming majority of PJM’s projected demand growth. Their requirements are unique: round-the-clock, firm, and non-negotiable. To ensure reliability for such concentrated loads, grid operators must procure significantly more standby capacity.
This surge in demand forecasts drove auction prices sharply higher. Because capacity costs are socialized, the extra burden landed on all customers — especially households. In effect, American families were helping subsidize Big Tech’s electricity appetite. This is the Hidden AI Tax in action.
Historical Trends and Market Dynamics
The story becomes clearer when viewed in historical context.
2018–2021: Capacity prices were relatively stable, often below $100/MW-day in PJM.
2022–2024: Prices dipped to historic lows, with one auction clearing at under $30/MW-day.
2025 onward: As data center growth accelerated and more aging coal and gas plants retired, auction prices skyrocketed nearly tenfold.
Other drivers compounded the trend. Interconnection queues for new renewable and storage projects created delays in bringing fresh supply online. Market rule changes tightened the accreditation of intermittent sources like solar and wind, requiring more “firm” capacity to meet reserve margins. In New England, winter fuel security concerns also added to costs.
The players in this market are diverse: Big Tech companies demanding new supply, generator owners retiring plants or bidding higher, regulators like PJM and ISO-NE balancing reliability, and consumers ultimately footing the bill. The winners are capacity sellers and investors; the losers are households with inflated electricity bills.
India’s Power Sector: A Different Model
Could India face a similar fate? The short answer: possible, but not inevitable.
India’s electricity sector is structured differently. Power distribution is handled largely by state-owned DISCOMs (distribution companies). These entities purchase electricity through long-term power purchase agreements (PPAs) with generators, often spanning 20–25 years. Prices are regulated by state electricity commissions (SERCs) under central guidelines (CERC).
A defining feature of India’s tariff system is cross-subsidy. Industrial and commercial customers typically pay higher tariffs, while households and agricultural users receive subsidized rates. In contrast to the U.S., where households are subsidizing industrial data center growth, India’s system shifts more burden onto industry to shield domestic consumers.
Moreover, India does not yet have a mature, nationwide capacity auction market like PJM. Capacity payments exist but account for only a modest share of overall costs — estimated at around 10 percent. Policy discussions are ongoing about introducing capacity mechanisms to ensure reliability as renewable penetration increases.
India’s Data Center Boom
India’s situation is changing rapidly. Data center capacity in India has grown from about 350 megawatts in 2019 to over 1,000 MW by 2024, and projections suggest it could reach 1,800 MW by 2026. Key hubs include Mumbai, Navi Mumbai, Hyderabad, Noida, Chennai, and Bengaluru.
With government incentives, the rollout of 5G and cloud services, and localization policies, India is positioning itself as a data center and AI hub. This raises a critical question: will India follow the U.S. path, where household bills are inflated by industrial demand, or can it chart a different course?
Power Grid Corporation of India Limited: The Backbone
One of India’s strengths lies in its robust transmission backbone, operated by Power Grid Corporation of India Limited (PGCIL).
PGCIL owns and operates over 176,000 circuit kilometers of high-voltage transmission lines and more than 400 substations, covering around 85 percent of India’s inter-state transmission capacity. With system availability consistently above 99.7 percent, PGCIL ranks among the most reliable operators globally.
Crucially, through projects like the Green Energy Corridors, PGCIL is enabling the integration of over 20 gigawatts of renewable capacity into the national grid. This infrastructure allows renewable power from Rajasthan’s deserts or Tamil Nadu’s coasts to be transmitted efficiently to urban and industrial centers, including future data hubs.
By ensuring both reliability and sustainability, PGCIL acts as a safeguard against the kind of localized shortages and bottlenecks that have driven U.S. capacity auction prices so high.
Implications for Indian Youth
For India’s youth, the Hidden AI Tax story offers several lessons and opportunities.
First, it highlights the importance of energy economics as a policy choice. Electricity tariffs are not just technical matters but political decisions about who pays and who benefits. An informed generation can demand transparency and fairness.
Second, it signals career opportunities. The rapid transformation of India’s power sector is opening paths in smart grid engineering, renewable integration, AI-driven energy optimization, and policy analysis.
Third, it calls for activism. Youth voices in public hearings, petitions, and regulatory consultations can influence how new capacity costs are allocated. If citizens remain silent, households may eventually bear the burden, as in the U.S.
Finally, it invites entrepreneurship. Startups focused on rooftop solar, battery storage, demand response, and renewable solutions for data centers can play a crucial role in shaping a fairer digital economy.
Toward Optimistic Solutions
The good news is that India still has a chance to design a better model. Several solutions stand out:
- Mandating renewable energy for data centers: Ensure hyperscale loads procure 100 percent renewable-backed supply through long-term contracts.
- Fair tariff design: Allocate incremental capacity costs directly to large industrial consumers rather than households.
- Smart grids and storage: Scale up battery systems and demand response to avoid auction price spikes.
- Transparent capacity mechanisms: If India adopts capacity markets, they should be designed with safeguards ensuring that new large loads bear their share of the costs.
- Youth participation: Encourage young Indians to shape the future through careers, activism, and entrepreneurship in the energy sector.
Conclusion: Who Pays for the Digital Future?
The rise of AI is reshaping not just technology but also energy systems. In the United States, the cost of powering data centers is already being passed onto households in the form of inflated electricity bills. The Hidden AI Tax is a warning signal.
India has the advantage of hindsight. With strong transmission infrastructure, regulated tariffs, and cross-subsidies, it is better positioned to shield households. Yet with the rapid growth of AI and data centers, complacency could still lead to similar pitfalls.
Ultimately, the choice lies with policymakers and the rising generation. The digital future will certainly demand more power. The real question is: who will pay the bill?
No comment yet, add your voice below!