How is Google Balancing Clean Energy Grids & AI Demand?

Although AI-driven transformation promises innovation and economic opportunity, the colossal challenge of data centres’ massive, ever-growing appetite for electrical power continues to rear its head.

The International Energy Agency (IEA) indicates that in the United States, power consumption by data centres is set to account for almost half of the growth in electricity demand from now until 2030, primarily driven by AI advancements.

Tech giant Google is leading strides in AI and sustainability.

No stranger to creating tech-driven solutions, it is leveraging this position to tackle the AI power challenge by making its data centres smarter and more flexible than ever before.

How Google is rethinking data centre power

Typically, grid operators maintain excess power capacity to ensure supply during peaks in demand. In the US, only about 50% of generating capacity tends to be utilized on average, while the remainder is reserved for high-demand periods. Constructing new generation or transmission infrastructure to accommodate large, inflexible AI and ML workloads can prove both costly and time-intensive.

This is where Google steps in. It has reimagined the approach, asking: why not make energy demand itself more adaptable instead of modifying the grid for ever-larger peaks?

Such insights have spurred the development of demand flexibility — a technique that shifts or reduces energy use in real time, especially concerning large computing operations. By synchronizing compute demand with periods of abundant, clean energy or lower grid strain, Google can support integrating AI into the economy without heightening power system constraints.

“Innovation isn’t just about developing brand new shiny things. In fact, some of the most important innovations come from collaborations to make existing systems more intelligent — and in this case, more flexible," says Google’s Chief Sustainability Officer, Kate Brandt.

“We’re sharing our advancements with new flexible demand capabilities in our data centres, now for the first time by targeting ML workloads. 

“This new approach can support AI growth and our grid partners at the same time — helping utilities reliably and cost-effectively meet the electricity needs of all their customers.

“While this is still early stages, we see demand response as a promising tool to shift or reduce our power demand, providing flexibility when the grid needs it most.”

The power of demand response

Demand response, where electricity consumers like Google adjust their power usage in response to incentives or price signals, typically during peak demand or when renewable energy output is low, is crucial, according to the IEA. Technologies like demand response play a significant role in lessening the requisite for "costly new transmission and distribution infrastructure”.

Demand response is now capable of targeting machine learning workloads — the computational nucleus of modern AI — to assure data centre clusters are more adaptable. Contingently, Google has collaborated extensively with Indiana Michigan Power (I&M), Tennessee Valley Authority (TVA) and Omaha Public Power District (OPPD), commencing the large-scale implementation of these capacities.

In a pilot scheme with OPPD, Google demonstrated the capability to curtail ML-related power demand amid grid stress, propelling the path for broader utilisation with new utility partners.

“I&M is excited to partner with Google to enable demand response capabilities at their new data centre in Fort Wayne, IN,” says Steve Baker, President and Chief Operating Officer of I&M.

“As we add new large loads to our system, it is critical that we partner with our customers to effectively manage the generation and transmission resources necessary to serve them.

“Google’s ability to leverage load flexibility as part of the strategy to serve their load will be a highly valuable tool to meet their future energy needs.”

Flexible demand benefits both AI and grids

By rendering massive machine learning demand more flexible, Google facilitates several key benefits:

• Faster AI deployment: Data centres can be integrated into the grid more swiftly, without waiting for new power plants or transmission lines, expediting AI service availability.
• Lower costs and carbon: Utilities satisfy new demand through existing infrastructure, minimizing the need for capital-heavy projects and decreasing the carbon footprint.
• Grid resilience: Real-time load management enhances the overall robustness of the power system, a necessity in an era of increasing renewable energy and variable supply.

Google’s 24/7 carbon-free energy objective — matching each hour of electricity use with clean energy — has catalysed the creation of technologies and strategies that foster its sustainability goals and assist the broader grid.

Smoothing out electricity peaks and valleys, flexible scheduling of compute tasks for when clean power is plentiful, and avoiding peak periods make renewables easier to manage and waste reduction more feasible.

“As AI adoption accelerates, we see a significant opportunity to expand our demand response toolkit, develop capabilities specifically for ML workloads and leverage them to manage large new energy loads,” Michael Terrell, Google’s Head of Advanced Energy, says.

“By including load flexibility in our overall energy plan, we can manage AI-driven growth even where power generation and transmission are constrained. 

“We believe this is a promising tool for managing large new energy loads and facilitating investment and growth.”

Evolving to keep up with AI and data centre demand

Data centre demand flexibility is confined to select locations, but Google persistently believes extending flexibility to include ML tasks signifies a crucial advancement towards wider demand management, reinforcing grid stability and cost efficiency.

“Incorporating ML workloads is an important step to enable larger scale demand flexibility, delivering grid reliability and cost-saving benefits in the places where these capabilities are deployed,” Michael adds. 

“By engaging in long-term resource planning with utility partners like I&M and TVA, we can integrate flexibility into future grid development alongside Google’s data centre infrastructure deployment.”

He maintains that addressing the burgeoning energy demands of data centres necessitates diverse strategies — and demand response stands out as an essential tool.

“Looking forward, we remain committed to collaborating with system operators, utilities and industry partners to capture AI’s immense opportunity while supporting a clean, reliable and affordable energy system for everyone,” he concludes.