Decoding The Nexus of Climate, Data, and Energy
Written by Harshit Agrawal, MEM '26
Decoding the nexus of climate, data, and energy may be the defining challenge of the modern era. I explored this intersection across a 9-article series examining key dimensions of the power demand surge driven by data centers.
The scale of demand is staggering: according to the IEA, global electricity use by data centers is projected to hit 945 TWh annually by 2030, more than double the 415 TWh in 2024. In the U.S., the number of hyperscale data centers rose from 147 in 2015 to over 550 by 2025, with their capacity nearly doubling. The interplay between data centers and renewables is complex: on one hand, data centers' energy hunger is driving investment in renewables and could make emerging green technologies scale faster; on the other hand, if not managed carefully, it can also lock in new fossil fuel infrastructure (as a "quick fix" for reliability) and make it harder to decarbonize the grid. Even if we solve generation, interconnection remains a major hurdle; more than double the current U.S. capacity is stuck in queues. Upgrading the grid is highly cost-intensive for utilities. The economic burden haunts regulators and, by extension, the general public, who might end up bearing the brunt of upgrade costs driven by data centers. As a result, this becomes a policy (externality) question that regulators across the globe are solving differently. For example, Denmark mandates waste heat reuse, Singapore limits builds via permit quotas, and Virginia is rethinking tax incentives. With all of this, and the cutthroat competition to take the AI lead, hyperscalers are doing all they can to source reliable and cheap power, even if that means setting up independent grids, investing in alternative generation (nuclear, geothermal, etc.), allying with peers and the grid (like Google–PJM), and supporting the grid with the same AI that's straining it in the first place. On the solutions front, hyperscalers are joined by emerging startups that are trying to add value in this complex but crucial ecosystem, solving one problem at a time.
Ultimately, data centers becoming grid-friendly and green-friendly is a question of time. All eyes on AI.