IT Brief Australia - Technology news for CIOs & IT decision-makers
Story image

Exclusive: AWS’ sustainable cloud push drives 99% carbon footprint reduction for Aussie users

Thu, 7th Nov 2024

Amazon Web Services (AWS) is making significant strides in helping businesses reduce their carbon emissions through cloud migration.

Jenna Leiner, ESG and external engagement lead at AWS, shared insights from a recent study conducted with Accenture, demonstrating the substantial sustainability benefits that come from moving AI workloads to the AWS cloud.

"This is the second study we've done in this space," Leiner said, during an interview.

The study focused on comparing AI workloads on AWS' infrastructure to those on traditional on-premises data centres, revealing that AWS is, on average, 3.2 times more energy efficient. In Australia, the impact is even more pronounced, with businesses potentially reducing their associated carbon footprint by up to 99% when optimising on AWS.

Leiner explained the methodology behind the study: "We simulated a compute-heavy and a storage-heavy workload on AWS and then replicated that same workload on on-premises data centres. Accenture's analysis showed that the gains come from a combination of more efficient hardware, improved power and cooling efficiency, and AWS' ability to procure carbon-free energy at scale."

The Benefits of Migrating to AWS

The 'lift and shift' approach, where workloads are transferred to AWS without changing their underlying architecture, is key to achieving sustainability gains. According to Leiner, "By lifting and shifting a workload from on-prem to AWS, you can reduce the associated carbon emissions by 94% for operational and embodied hardware emissions."

Optimisation within AWS takes these reductions even further.

"That incremental lift to 99% comes when you optimise. For example, adopting our custom silicon, such as AWS Graviton processors, can improve both performance and energy efficiency," she explained.

Leiner described the deeper efficiency gains AWS has achieved through innovative design choices. "We removed the central uninterruptible power supply (UPS) from our data centre design, replacing it with small battery packs integrated into each rack. This change reduced energy conversion loss by 35%," she said, underlining AWS' commitment to sustainable practices.

Specific AI Optimisations

AWS' tailored solutions for AI and machine learning (ML) workloads further enhance energy savings. Leiner detailed the impact of AWS Trainium, a machine learning chip designed to cut training times significantly. "Trainium reduces training time from months to hours, translating to cost savings and up to 25% reductions in energy consumption compared to similar solutions," she said. These advancements are crucial as more businesses experiment with AI, driving demand for powerful yet sustainable computing solutions.

Tackling the Barriers to Transition

Transitioning to cloud-based solutions is not without challenges. Leiner, who previously worked with large enterprises on digital transformation, noted that "the barriers are more cultural than technological."

"It's about aligning the CIO, CTO, and other key leaders to support the transition," she said. AWS assists customers by developing tailored business cases, projecting cost profiles, migration journeys, and potential carbon emission reductions. "A common misconception is that simply moving data to the cloud is enough. Instead, it's essential to migrate strategically, focusing on data that will drive efficiency or optimisation," she added.

AWS has seen encouraging feedback from customers who have successfully made the transition. Leiner cited Canva as a notable example, saying, "Canva saw a 90% reduction in their carbon emissions by running their workloads on AWS." Global health company Illumina also reported an 89% reduction, demonstrating the far-reaching impact of cloud migration.

AWS' Global Sustainability Efforts

Beyond individual customer successes, AWS is committed to sustainability at a broader scale. Leiner shared that AWS achieved its renewable energy goal seven years ahead of schedule, matching the electricity used in its global operations with 100% renewable energy in 2023. "We are also exploring new methods such as liquid cooling, which could enhance power and water efficiency in our data centres," she said.

AWS' approach to innovation is comprehensive. "Every new basis of design for our data centres globally prioritises energy efficiency," Leiner emphasised. She pointed to AWS' scale and commitment as a significant advantage. "Because of our size, we can test boundaries and push innovation that leads to meaningful, scalable changes in sustainability practices," she said.

A Future Driven by AI and ESG

AI's potential role in sustainability and Environmental, Social, and Governance (ESG) strategies is promising.

"It's early days, but AI has the potential to address complex challenges related to climate change that we haven't solved before," Leiner said. She pointed out that AWS' focus is not solely on leveraging generative AI but on ensuring that the most suitable technologies are used for each challenge.

"Our teams focus on understanding what problems need solving and then using the best tools to do so," she explained.

This balanced approach means that AWS' investments in custom silicon and energy-efficient infrastructure will continue to support companies as they scale their AI operations sustainably.

"The shift in attitude among Australian IT teams, now considering sustainability alongside cost and resiliency, is inspiring. It's fantastic to see the progress we're making."

Follow us on:
Follow us on LinkedIn Follow us on X
Share on:
Share on LinkedIn Share on X