banner

News

Oct 16, 2024

Data center growth demands long-duration energy storage | Trellis

The key to unlocking a future where our insatiable energy demand is fulfilled by the lowest-cost and cleanest form of generation. Read More

This article is sponsored by Hydrostor.

The expansion of our electricity grid to support data center growth is at the forefront of the energy debate. Even relying on traditional energy sources, this would be a challenge, but doing it in a way that prioritizes decarbonization is going to be even tougher.

A ChatGPT query requires nearly 10 times as much electricity to process as a Google search, and AI services that require immense power are being used more and more often by Americans.

Data centers operated by Amazon, Microsoft, Google and Meta (the “hyperscalers”) already represent 9 gigawatts (GW) of capacity in the United States alone—and this is projected to triple by 2028, reaching nearly 28GW.

The trends are clear. Virginia, also known as the “data center capital of the world,” is expected to exceed 10GW of electricity demand from data centers by 2028, and Amazon is planning upwards of US$35 billion of investment by 2040 to support expansion across the state.

Texas has also been identified as a prime market for hyperscalers, facing immense pressure with more than 70 percent of new interconnection requests coming from data centers, and The Electric Reliability Council of Texas (ERCOT) estimating peak load demand to exceed 152GW by 2030.

Beyond Virginia and Texas, independent electricity system operators (IESOs) are facing unprecedented demand growth and straining transmission infrastructure, while simultaneously curtailing excess renewable power.

Most markets with high renewable penetration lack storage resources with durations longer than four hours to time-shift vast amounts of solar for serving data centers’ 24/7 loads. This reality only becomes more challenging when paired with decarbonization objectives driving more renewables onto the grid.

Do we stall the push for decarbonization and continue to delay coal retirements or adding new natural gas resources to ensure adequate generation and supply to meet the growing demand of AI and electrification? Or is there a way to maximize the renewable energy we’re adding to the grid without jeopardizing reliability or the future of our planet?

Long-duration energy storage could be the answer to the problem, and it will be critical to maintain the widescale deployment of renewable energy sources. Long-duration energy storage, or LDES, is the key to unlocking a future where our insatiable energy demand is fulfilled by the lowest-cost and cleanest form of generation—wind and solar.

LDES is distinct from traditional energy storage deployments like lithium-ion batteries and refers to technologies that can store energy for periods of 8 hours and beyond, encompassing a range of approaches that vary in their maturity and commercial readiness.

Hydrostor, founded in 2010, is one of the few LDES companies whose technology is commercially ready. Hydrostor has raised about US$350 million from the likes of Goldman Sachs and Canadian Pension Plan Investments. To date, Hydrostor has executed more than US$2 billion of energy storage power purchase agreements (PPAs).

As a developer and operator of its facilities, Hydrostor has reinvigorated an age-old approach to LDES in the form of Advanced Compressed Air Energy Storage (or A-CAES, pronounced A-“CASE”).

Hydrostor’s technology works by taking renewable energy from the grid when wind and solar might otherwise be curtailed and storing that energy in a cavern deep underground filled with water, in the form of hydrostatic pressure. When energy is needed, that pressure is released, spinning turbines, generating electricity.

Offering distinct advantages when compared to alternative energy storage technologies, A-CAES has a 50+ year lifetime, zero efficiency degradation, is flexible to locate, has no fire risk, provides spinning reserves (inertia is not found with inverter-based technologies like batteries, solar and wind) and doesn’t rely on hard-to-recover or imported minerals.

A-CAES offers operational flexibility for a changing grid and utility needs with no requirements around depth of discharge or state of charge. It can be fully depleted and fully charged with no degradation or cycling limitations while other technologies experience accelerated degradation if not operated within charging parameters. But most important, it is deployable today.

Hydrostor’s patented A-CAES solution is unique because it uses components from traditional mining and gas operations to provide a low-impact, low-cost, long duration energy storage solution. This also means that it is highly scalable and suited for utility scale deployments that exceed hundreds of megawatts for an individual project.

The 500MW, 8-hr Willow Rock Project in California and the 200MW, 8-hr Silver City Project in New South Wales are Hydrostor’s most advanced projects, both slotted to start construction in 2025. With a 7GW pipeline globally, in addition to these late-stage projects, A-CAES is well-positioned to play a key role fulfilling data center electricity demand with clean power.

The growth of Hydrostor’s pipeline is due to the successful commissioning and operation of the Goderich Energy Storage Center. Located west of Toronto, this project is contracted by the local IESO to provide grid services and conforms to all interconnection, uptime, performance and dispatch requirements.

Data center owners and operators can be empowered to feed their voracious appetite for electricity at all hours of the day and night, without taking a significant toll on the grid, or the environment—but incorporating more LDES onto the grid is the key.

By proactively embracing LDES solutions, data centers and utilities can not only ensure their own operational resilience but also play a pivotal role in building a stable and sustainable grid. This symbiotic relationship is essential for managing rapidly increasing electricity demand while simultaneously meeting decarbonization goals.

SHARE