AWS partners with Orbital Materials to boost carbon removal, cooling, and efficiency in data centers

Data center racks with cables and servers

Inside a data center where storage servers populate dozens of racks (Image credit: Old Source)

AWS and Orbital want carbon-negative data centers by end of 2025
Carbon removal costs are estimated at $0.20 per GPU hour
Orbital’s AI model Orb will debut on AWS platforms

AWS and Orbital Materials have announced a partnership using generative AI to boost data center sustainability and efficiency.

The partnership will focus on developing new materials for carbon removal, chip cooling, and water utilization within AWS data centers.

Traditional material development has relied on lengthy trial-and-error processes in laboratories, but Orbital Materials says its own generative AI platform can allow rapid design, synthesis, and testing of advanced materials.

Generative AI transforms materials discovery for faster, efficient innovation

Since establishing its lab in early 2024, Orbital says it has achieved a tenfold improvement in material performance, showcasing unprecedented speed in materials discovery.

The company's first product, a proprietary carbon removal material for direct air capture (DAC) has captured AWS's interest, allowing Orbital to develop a DAC system designed to operate within data centers, repurposing waste heat to drive carbon removal, and significantly reducinges scope 3 emissions, aligning with AWS’s sustainability goals.

Data centers are critical to powering modern technology but are also significant energy consumers with complex carbon footprints. Through the partnership, AWS aims to integrate Orbital’s DAC system into its data centers, demonstrating a full-scale pilot by the end of 2025.

Orbital estimates its material could achieve carbon negativity for renewable-powered data centers for $0.20 per GPU per hour, a small fraction of current GPU usage costs, which range from $2 to $10 per hour. The potential economic viability of this technology could pave the way for scalable adoption across the cloud computing industry, enabling AWS and others to mitigate environmental impacts.

As part of the partnership, Orbital’s generative AI model, Orb, will become available to AWS customers through Amazon SageMaker JumpStart and AWS Marketplace. Orb is the first AI-for-materials model on AWS platforms and will enable researchers and businesses to simulate and develop advanced materials more efficiently. Applications for Orb extend beyond carbon removal and include technologies like semiconductors, batteries, and electronics.

To support its AI advancements, Orbital will pre-train and fine-tune its foundation models on Amazon SageMaker HyperPod, AWS’s purpose-built infrastructure for large-scale distributed training. Additionally, Orbital plans to evaluate the deployment of AWS Trainium, a custom silicon designed to optimize the cost performance of deep learning workloads.

“Our partnership with AWS will accelerate the deployment of our advanced technologies for data center decarbonization and efficiency. Working with the market-leading AWS team will ensure that our suite of products in cooling, water utilization and carbon removal enables the next generation of data centers powering the AI revolution,” said Jonathan Godwin, CEO and Co-Founder of Orbital Materials

Howard Gefen, General Manager of AWS Energy & Utilities said "AWS looks forward to collaborating with Orbital and their mission to drive data center decarbonization. Through Amazon SageMaker HyperPod and AWS Trainium, we can accelerate the development of breakthrough sustainability technologies."

"By integrating Orb with Amazon SageMaker JumpStart and AWS Marketplace, we will enable sustainable innovation more widely. Together, we have the opportunity to set new benchmarks for carbon removal and efficiency across the industry," Gefen concluded.

Efosa has been writing about technology for over 7 years, initially driven by curiosity but now fueled by a strong passion for the field. He holds both a Master's and a PhD in sciences, which provided him with a solid foundation in analytical thinking. Efosa developed a keen interest in technology policy, specifically exploring the intersection of privacy, security, and politics. His research delves into how technological advancements influence regulatory frameworks and societal norms, particularly concerning data protection and cybersecurity. Upon joining TechRadar Pro, in addition to privacy and technology policy, he is also focused on B2B security products. Efosa can be contacted at this email: udinmwenefosa@gmail.com