SparkNotes
Posts
SparkNotes: AI and the Data Center Boom

SparkNotes: AI and the Data Center Boom

A Guest Post with Alan Cordova, Senior Advisor at Ever.green

Julia Wu & Alan Cordova
October 23, 2024

Dear Spark community,

To those we met in person at RE+ in September, thank you for participating in Spark’s first-ever booth exhibition. We experienced incredible foot traffic and enjoyed seeing new and familiar faces.

In this special October issue of SparkNotes, Spark advisor Alan Cordova shares his perspective on the ongoing data center boom. Alan is a Senior Advisor of Renewables at Ever.green (a leading tax credit marketplace), and formerly a VP of Business Development at Fluence (global market leader in energy storage products and services).

Our team is en route to London to participate in Amazon’s GenAI Day for Energy & Utilities. There, we will join 14 other startups selected from 800 candidates worldwide to engage with AWS corporate customers in the energy sector. We look forward to sharing what we learn in the next issue!

AI and the Data Center Boom

In the past few years, data centers have become the most in-demand asset in the power industry. Demand for computing power has skyrocketed as energy-intensive generative AI tools have started to achieve widespread adoption. According to Goldman Sachs, these tools are much more energy-intensive than their predecessors - a ChatGPT query needs nearly 10 times as much electricity to process as a Google search - and, consequently, data centers are forecasted to overwhelm incremental gains in energy efficiency achieved through the evolution of the computing technology. McKinsey forecasts that data center power consumption will rise from 147 TWh / year in 2023 to 606 TWh / year by 2030.

Data center developers (and their end-use customers, such as Amazon) prioritize scale, execution risk, cost, speed to market, reliability, and access to clean energy. These priorities often compete, and developers are often forced to sacrifice one or more, invest in new technologies or solutions, or seek partnerships with other stakeholders. In many instances, data centers needing reliable round-the-clock energy quickly sacrifice environmental goals by pursuing on-site generation fueled by natural gas. More recently, data center owners have sought to restart decommissioned nuclear plants, supporting environmental goals while taking significant execution risk.

According to SeekingAlpha, nuclear energy has advantages to data center owners over wind or solar: nuclear creates a larger amount of power that is stable and consistently available on demand. Recently, Amazon, Google, and Microsoft announced $500M of investments in nuclear energy to satisfy its data center demands and artificial intelligence ambitions while continuing to meet its clean energy goals. Amazon is signing a deal with Dominion Energy to explore building a small modular nuclear reactor near the North Anna station. A few days before, Google announced an agreement to purchase power generated from multiple small modular reactors to be built by nuclear energy startup Kairos Power. Microsoft has also teamed up with Constellation Energy to resurrect a unit of the Three Mile Island nuclear power plant in Pennsylvania.

Beyond tech giants, many data center developers seek grid connections that enable them to access off-site generation through a power purchase agreement; such developments subject them to lengthy and uncertain interconnection processes administered by a grid operator (such as PJM in the mid-Atlantic region) that is overwhelmed by similar requests on both the demand and supply side. Operators have warned that data centers may face delays of up to seven years before they are allowed to connect to the grid. Identifying sites with available withdrawal capacity, or where capacity may become available more quickly, has become a competitive advantage for developers.

The issues confronting data center developers are familiar to renewable energy and energy storage developers. Data center development follows very similar pathways to power generation development, with the same basic set of stage gates and objectives (e.g., land entitlement, permits, grid interconnection, power purchase agreement) needed before construction begins. Of critical concern is the local impact: just as solar and wind projects face viewshed and habitat concerns due to their physical footprint, data centers must contend with their sonic impact and water consumption. Similar skills and tools are needed to screen sites, establish priorities, and advance projects.

The challenges facing data center developers strikingly mirror those in renewable energy development:

Complex permitting processes
Uncertain grid interconnection procedures
Balancing environmental goals with reliability and cost
Managing community engagement and local impacts

This parallel highlights the increasing convergence of energy-intensive industries and the need for sophisticated solutions to navigate their complexities.

Source: Goldman Sachs

As the energy landscape evolves, with data centers, renewable energy, and emerging technologies becoming increasingly intertwined, Spark is positioned to play a meaningful role. By leveraging our expertise in complex regulatory environments and optimizing development workflows, we aim to support the renewable energy sector as well as the broader ecosystem of energy-intensive industries.

If you are currently developing front-of-the-meter renewable energy assets, we would love to hear from and support you! Book a demo.

Further Reading