Google's Semiconductor Innovation Reduces AI Footprint
Google has published an eye-opening study examining the complete lifecycle emissions of its AI accelerator chips.
“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 agrees, adding: “This is just the beginning with huge opportunities to continue optimising hardware and software for carbon efficiency.”
Google's approach: Measuring Compute Carbon Intensity
As part of this study, Google has introduced Compute Carbon Intensity (CCI), a new metric designed to enhance transparency and drive innovation across the industry.
The research analysed five models of Tensor Processing Unit (TPU) hardware, assessing their full lifecycle emissions to explore how design choices influence carbon efficiency.
TPUs are specialised hardware accelerators that play a crucial role in advancing AI.
“Their efficiency impacts AI's environmental sustainability," explains Robert Little, Sustainability Strategy Lead for gTech at Google.
"This progress is due to more efficient hardware design, which means fewer carbon emissions for the same AI workload.”
CCI quantified an AI accelerator chip’s carbon emissions per unit of computation – this is measured in grammes of CO₂ per Exa-FLOP.
A lower CCI score therefore means lower emissions from the hardware platform for a given AI workload.
Google has used CCI to track its progress in increasing the carbon efficiency of its TPUs.
Insights from Google’s research
Google’s research found a threefold improvement in the CCI of its TPU chips over four years. This spans the evolution from TPU v4 to Trillium.
By adopting newer generations of TPUs, its customers produce fewer carbon emissions for the same AI workload.
Operational electricity emissions make up more than 70% of a Google TPU’s lifetime emissions, highlighting the importance of improving the energy efficiency of AI chips. It is also vital to reduce the carbon intensity of the electricity powering these chips.
While operational emissions account for the majority of a TPU’s lifetime emissions, manufacturing emissions still play a role and their share of total emissions will increase as operational emissions decline.
Google’s detailed manufacturing LCA helps it focus its decarbonisation efforts on the highest-impact initiatives and it is actively working with its supply chain partners to reduce these emissions.
“These findings highlight the importance of optimising both hardware and software for a sustainable AI future,” Robert concludes.
“It's important to remember where AI has important implications for reducing emissions and fostering sustainability.”
Explore the latest edition of Manufacturing Digital and be part of the conversation at our global conference series, Manufacturing LIVE.
Sign up to our weekly newsletter here.
Discover all our upcoming events and secure your tickets today.
Manufacturing Digital is a BizClik brand.
