How Google’s AI Chips are Boosting Chip Carbon Efficiency
Google is shedding light on the entire lifecycle emissions of its AI accelerator chips with a first-of-its-kind study, marking a significant stride in the tech giant's sustainability efforts.
“The study found that innovation in our chip hardware design led to a 3x improvement in the carbon-efficiency of AI workloads over two generations and that decarbonising our electricity-related emissions will drive the biggest carbon reductions for our AI footprint,” explains Kate Brandt, Chief Sustainability Officer at Google.
“At Google, we know AI can drive transformative innovation in areas like information, optimisation and prediction.
“We also know it’s equally important to manage its environmental impacts, and we’re working to do that through efficient infrastructure, model optimisation and emissions reductions.
“This study is an important step in those efforts, unlocking critical insights for Google and others looking to reduce emissions across the full lifetime of AI hardware.”Adam Elman, Director of Sustainability EMEA at Google, says: “This is just the beginning with huge opportunities to continue optimising hardware and software for carbon efficiency.”
Introducing Compute Carbon Intensity
A key outcome of the study is the introduction of the Compute Carbon Intensity (CCI), a novel metric that aims to bring enhanced transparency and encourage greater innovation industry-wide.
Google's research involved a close examination of five versions of its Tensor Processing Unit (TPU) hardware, helping to gauge the full lifecycle emissions and understand how specific hardware design choices affect overall carbon efficiency.
TPUs, specialised hardware accelerators, are pivotal to advancing AI technologies and their ongoing sustainability.“Their efficiency impacts AI's environmental sustainability. This progress is due to more efficient hardware design, which means fewer carbon emissions for the same AI workload,” explains Robert Little, Sustainability Strategy Lead for gTech at Google.
He describes the CCI metric as a measure of an AI accelerator chip’s carbon emissions per unit of computation, quantified in grams of CO₂ per Exa-FLOP — a lower CCI score indicates reduced emissions for a particular AI workload.
Through this innovative CCI metric, Google has been able to monitor and enhance the carbon efficiency of its TPUs effectively.
Unpacking the findings
The company’s findings are nothing short of transformative: a tripling of the CCI for its TPU chips across four years, from the TPU v4 to the Trillium models.
This improvement means Google's customers can now achieve lower carbon emissions for equivalent AI workloads, simply by opting for newer TPU generations.
Digging deeper, operational electricity emissions make up more than 70% of a Google TPU's total lifecycle emissions, highlighting the importance of targeting energy efficiency improvements.
Furthermore, as these operational emissions are scaled back, the proportion of emissions attributed to manufacturing begins to grow.
This insight has led Google to focus its decarbonisation strategies in manufacturing areas likely to yield the highest impacts, facilitated by a detailed manufacturing lifecycle assessment (LCA).
“These findings highlight the importance of optimising both hardware and software for a sustainable AI future,” Robert says.
“It's important to remember where AI has important implications for reducing emissions and fostering sustainability.”
Explore the latest edition of Energy Digital Magazine and be part of the conversation at our global conference series, Sustainability LIVE.
Discover all our upcoming events and secure your tickets today.
Energy Digital is a BizClik brand
