Terrill Dicki
Dec 02, 2025 17:05
NVIDIA and AWS have expanded their collaboration, integrating new technologies to enhance AI infrastructure and cloud capabilities, promising more efficient, scalable solutions for global enterprises.
NVIDIA and Amazon Web Services (AWS) have announced an expansion of their strategic partnership, focusing on advanced technology integrations aimed at boosting AI infrastructure and cloud capabilities. This development was unveiled at the AWS re:Invent conference, highlighting the integration of NVIDIA’s NVLink Fusion with AWS’s custom silicon, including the next-generation Trainium4 chips. This collaboration is expected to enhance performance and accelerate the deployment of cloud-scale AI capabilities, according to NVIDIA’s official blog.
Technological Advancements
The expanded partnership will see AWS supporting NVIDIA NVLink Fusion, a platform designed for custom AI infrastructure. This support will facilitate the deployment of AWS’s custom-designed silicon, such as Trainium4 chips for AI model training, and Graviton CPUs for various workloads. The integration aims to increase performance and reduce time-to-market for AWS’s cloud-scale AI capabilities.
Additionally, AWS plans to integrate Trainium4 with NVIDIA’s MGX rack architecture, marking the beginning of a multigenerational collaboration between the two tech giants. AWS has already implemented MGX racks at scale with NVIDIA GPUs, and the integration of NVLink Fusion is set to streamline deployment and system management across AWS platforms.
AI Infrastructure and Global Reach
NVIDIA and AWS are committed to deploying sovereign AI clouds globally, bringing advanced AI capabilities to various industries while meeting local regulatory requirements. The launch of AWS AI Factories will provide customers with dedicated infrastructure for harnessing advanced AI services, allowing them to maintain data control and comply with regulations.
For public sector organizations, AWS AI Factories will transform the supercomputing and AI landscape, integrating AWS’s reliable cloud infrastructure with NVIDIA’s advanced GPUs and computing platforms.
Software and Developer Integration
The partnership also extends to software stack integration, with NVIDIA’s open models now available on Amazon Bedrock. This integration allows developers to build generative AI applications and agents at a production scale, offering high efficiency and accuracy.
The collaboration also introduces enhancements to Amazon OpenSearch Service, which now features serverless GPU acceleration for vector index building, powered by NVIDIA’s open-source library. This advancement promises up to ten times faster vector indexing, significantly boosting productivity for dynamic AI applications.
Physical AI and Robotics
NVIDIA’s Cosmos world foundation models are now accessible as microservices on Amazon EKS, supporting real-time robotics control and simulation workloads. This integration enables robotics companies to efficiently train and validate robots using NVIDIA’s simulation frameworks.
Leading robotics companies are utilizing these capabilities for diverse applications, from data processing to scaling robotics development, showcasing the practical impact of the NVIDIA-AWS collaboration on the robotics industry.
This sustained partnership underscores NVIDIA’s recognition as the AWS Global GenAI Infrastructure and Data Partner of the Year, highlighting their joint efforts in advancing AI and cloud technologies.
Image source: Shutterstock
