Amazon Web Services (AWS) has unveiled new datacentre components that will ultimately be able to power more customer AI solutions.
- AWS IoT SiteWise Launches GenAI Industrial Assistant
- Amazon Ploughs $110M into Universities for AI Research
- Amazon Launches ‘Project Amelia’ AI-Powered Personal Assistant
By making updates to the power, cooling, and hardware designs within its datacentres, Amazon’s cloud computing platform is able to increase compute power by 12 percent and reduce energy consumption by 46 percent.
AWS’s datacentres support millions of customers around the world, including the hundreds of thousands of customers leveraging AI, machine learning services, as well as Amazon Bedrock, which can be used to build generative AI applications.
Prasad Kalyanaraman, Vice President of Infrastructure Services at AWS, commented on the new datacentre capabilities: “AWS continues to relentlessly innovate its infrastructure to build the most performant, resilient, secure, and sustainable cloud for customers worldwide.
“These datacentre capabilities represent an important step forward with increased energy efficiency and flexible support for emerging workloads.
“But what is even more exciting is that they are designed to be modular, so that we are able to retrofit our existing infrastructure for liquid cooling and energy efficiency to power generative AI applications and lower our carbon footprint.”
Datacentre Innovations
The simple fact is, as AWS points out, while the popularity in AI solutions has continued to grow as has the need for “increasingly higher power densities”.
As part of the solution, its latest datacentre design improvements therefore include simplified electrical distribution and mechanical designs.
For example, AWS has added backup power closer to the rack and reduced the amount of fans used to remove hot air, opting instead to use natural pressure differences to perform the task.
These will enable 99.9999 percent uptime and the simplified systems also cut down on the number of racks that can be affected by electrical issues by 89 percent.
AWS has developed mechanical cooling solutions for liquid-to-chip cooling in new and existing datacentres so that the most powerful AI chipsets like AWS Trainium2 can benefit from maximum performance and efficiency at the lowest cost.
By utilising data and generative AI, AWS has been able to optimise the way it positions its racks in datacentres so that it can make more efficient use of energy that is either unused or underused.
Further power has been unlocked through innovations to the power delivery systems that allow AWS to support a six times increase in rack power density for the next two years, and a further three times increase in the future.
From a sustainability perspective, on top of the 46 percent reduction in energy consumption, AWS has also slashed embodied carbon in the concrete of its datacentre building shell by as much as 35 percent.
Backup generators will also be run on renewable diesel that can lower greenhouse gas emissions by up to 90 percent across the fuel’s lifecycle.
James Bradbury, Distinguished Engineer, Compute, at Anthropic, shared its feedback on AWS’s datacentres: “As Anthropic develops our leading foundation models, having access to secure, performant, and energy-efficient infrastructure is crucial to our success.
“AWS’s commitment to building cutting-edge datacentres is one of the key reasons we’ve chosen them as our primary cloud provider and training partner.
“Their design improvements represent a significant step forward in providing secure, scalable, and efficient infrastructure to power AI models and drive innovation in this field.”
The latest datacentre components have been built to scale throughout AWS’s infrastructure around the world, with construction on AWS datacentres in the United States in early 2025.
