There is a lie that the tech world has repeated so often that the cases seem true: the digital economy is light, efficient and even ethereal. Moving data is less than moving atoms. “It’s just a click”, sounds in my head as I write these lines, recently sent to the head of the sector. That artificial intelligence, installed in it some abstract place called “the cloud”, does not have the common significance of a pipe, a pipeline or a coal-fired power station.
Empera, recent paper Economists Jamal Mamkhezri, Xiaochen Sol and Yuting Yang (Universidad Estatal de Nuevo México) were able to formalize what is obvious to anyone with two fronts: the cloud weighs and the electrical system notices its weight. Y see you again The balance is paid by someone who in most cases has no idea what they are doing.
These academics studied the event in Virginia between 2015 and 2024, the region with the highest density of data centers on the planet, and concluded that massive input into this infrastructure increases the local mayor’s clean electricity costs. It does not require the art of magic or a simple effect, but two mechanisms of equal or greater relief: red congestion and marginal transmission losses.
The combined effect is estimated at 7.3%. As Neil Armstrong once said, a small sum for a man; a big jump as you move up the ladder where the sector is growing.
We’ve added numbers to that price: The International Energy Agency, which doesn’t exactly know because of its alarmism, estimates that data centers will consume 415 TWh in 2024 (1.5% of global electricity) by 2025, and predicts that number will exceed 945 TWh in 2030. That’s more than the total electricity consumption of Japan, the third largest industrial economy on the planet.
Because what happened was a manageable externality, it was transformed into a vector of structural transformation of energy systems. Europe – which wants to triple its data processing capacity in six years after answering the European Parliament once or twice – is going down this road without having resolved the most fundamental question: who pays this red?
The unpleasantness of this topic is not rooted in basic economics, so much as what they imply politically. As the connection of data centers generates congestion and marginal losses that affect the highest price of electricity in their vicinity, the investigators describe mechanism of socialization of private costs.
Writing on the table: great technologies capture the benefits of massive data processing. The electrical system and possibly the domestic consumer absorb the friction. It is the oldest example of industrial capitalism applied to the infrastructure of the XXI century. century.
In Virginia, you were encouraged to succeed. In 2026, the state legislature debated legislation that would shift more red-line costs and capacity to large consumers — called data centers — with an explicit argument for reducing the goods invoice by four to five dollars per month. In PJM, North America’s main electricity market, the price of the last partial sale of capacity reached $329.17 per megawatt day, and the independent market watchdog estimated that demand for data centers accounted for roughly 40% of total costs. Cuarenta por cento. It is not an afterthought: it is the main force behind tariff building.
The tech industry’s usual argument against these critics has an elegance and a green halo you might recognize: “We buy renewable energy, we’re carbon neutral, AI will make us more efficient.” A good triumvirate of difficult arguments to answer. However, in this writer’s humble opinion It’s a scheme that every now and then fails at some crucial point.
Basic, technical: el paper shows that the damage to the electricity market does not come only from the volume of demand, until on geographical concentration and the inability of the transmissive red to absorb it without friction. A green energy contract does not eliminate congestion in a saturated node as you might imagine.
The second point is more political or communicative: the fact that a company declares carbon neutrality on its balance sheet does not solve the above question. here we are funding the red infrastructure necessary to ensure that it is physically neutral and not just countable.
Europe has an opportunity here, which it is likely to squander, as it usually does. On the old continent, we built a different regulatory identity in the digital space (privacy, data sovereignty, competence…) and now we have the opportunity to create a clear brand of how the energy costs of intelligence distribution are artificially distributed.
The IEA has made an explicit statement in this regard: the answer is not just to generate more electricity, but to flip it onto the grid, put infrastructure where real capacity is available and provide operational flexibility to large consumers. This requires regulation whether we like it or not. This requires political will, always politicians with sufficient altitude. But this is mainly because the public debate abandons the narrative of inevitable progress and the question remains that it is the same as everything before: Who exactly benefits from this revolution and who foots the bill?
Spain must be taught with more attention to what it shows in these moments, with the ministry on the horizon and the strategy in the field and in the arena with regard to data centers. Our country has seen huge data center projects come into being in recent years supported by the arguments of industrial modernization, European technological superiority and the creation of a skilled workforce. They are real arguments, but in every light incomplete. Because what Virginia has taught us is that by first bringing in this resolver-less infrastructure, transmission capacity, Red Costal separation, and local price impacts are obviously postpone the problem until it is more costly and politically sensitive to deal with it.
Leave a Reply