The Cloud's Crippling Cost

In climate community circles, it feels like everyone is talking about data centers and load growth, driven by the AI boom. Electric load growth in the U.S. has been anemic for years, rising a modest ~0.5% per year, but that may be about to change.

A December 2023 report from consulting firm Grid Strategies found that FERC filings in 2023 reflected a near doubling of grid planners’ five-year electricity load growth forecast, with “nationwide forecast of electricity demand [shooting] up from 2.6% to 4.7%.” The electric grid “is not prepared for significant load growth,” the report concludes.

Certain large data center hubs around the world have had to come up with new solutions to handle this strain on the grid - it is clear that flexible data centers are the solution and will soon become the new norm.

Case Study #1: Ireland

In the aftermath of the 2008 banking crisis, Ireland faced significant economic challenges. To recover, the government implemented policies aimed at attracting international technology companies, including offering exceptionally low taxes. The strategy worked and Ireland experienced a tech boom throughout the past 2 decades and has become a major data center hub for Europe. As of 2022, a fifth of all of the electricity used in Ireland is to power data centers (the same amount of electricity powers all urban households in the country). Projections indicate this could rise to one-third of total electricity by just 2026.

AWS is reportedly restricting the number of resources users can access in Ireland (Amazon's eu-west-1 region) amid ongoing concerns about the amount of power consumed by the nation’s data centers. AWS users informed The Register that there are often warnings that "You cannot spin up GPU nodes in AWS Dublin as those locations are maxed out power-wise.”

Case Study #2: Singapore

Singapore is a gateway to Asia for the rest of the world, and its high-quality infrastructure and wealth have helped it become one of the largest repositories for data storage and processing (Singapore has 60% of Southeast Asia's data market). Singapore decided to halt all new data center construction between 2019 and 2022 because of concerns that their high electricity demand would strain the grid and negatively impact the city-state’s sustainability goals. The pause lasted three years until the government called for applications in 2022 to build new facilities with higher environmental standards.

Both Singapore and Ireland have published rules in recent years that make new data center construction contingent on things like better energy and water efficiency, the ability to use backup generators, and “the ability to reduce power consumption when requested” (i.e., demand response). Instead of issuing a moratorium (like Singapore), Ireland now requires data centers to be dispatchable in exchange for a grid connection. So now every new data center in Dublin is getting a natural gas generator behind the meter so it can be flexible and avoid that contribution to system peak. (source)

New data center construction without additional, clean power risks raising power prices and grid emissions. We can expect more moratoriums if we get this next phase of data center growth wrong.

At Build AI, we see this writing on the wall and are only deploying data centers that use renewable energy and importantly have agreements with local utilities to not contribute to their peak loads (e.g. demand response). This means pausing AI models being trained at their checkpoints and turning off our data centers during the day (~5-8pm) when the grid is most strained, when power prices are the highest / the power is predominantly coming from fossil fuels. We’re a team of specialists that wants to make AI training less expensive, more accessible, and better for the planet.

Lower cost & lower environmental impact - start training your AI models with Build AI today. To get started, please complete our short questionnaire detailing your requirements.