NVIDIA and AWS Collaborate to Enhance AI Infrastructure with New GPU-Optimized Tools
Amazon Web Services (AWS) and NVIDIA have announced significant advancements in AI infrastructure, leveraging NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs and the NVIDIA cuVS library to improve performance, scalability, and cost-efficiency for enterprises.
What Are the Key Advancements in AWS’s AI Infrastructure?
AWS has introduced EC2 G7 instances powered by NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs, offering up to 4.6x faster AI inference performance compared to G6 instances. These instances are designed for production workloads requiring high performance without the operational overhead of a customer-managed GPU platform. The G7 instances support up to eight GPUs, 256GB of total GPU memory, and 700 Gbps of EFA-enabled networking, allowing customers to right-size infrastructure for their workloads instead of over-provisioning for them.
The NVIDIA cuVS library further accelerates vector search in Amazon OpenSearch Serverless, making GPU-powered indexing the default for all vector collections. This makes billion-scale vector databases practical to build in under an hour. Customers benefit from vector indexing up to 10x faster at a quarter of the cost, compared with CPU-only builds.
How Does AWS Achieve NVIDIA Exemplar Cloud Status for GB300?
AWS has achieved NVIDIA Exemplar Cloud status for NVIDIA GB300, meeting NVIDIA’s rigorous performance thresholds that it uses to benchmark AI workloads against its reference architecture. This designation ensures developers and AI leaders can be confident they’re using consistent, high-performance cloud infrastructure for large-scale training. The achievement is the result of deep co-engineering efforts between AWS and NVIDIA teams.
Why Does This Matter for AI Teams and Enterprises?
The advancements address critical challenges in AI deployment, including low-latency inference, scalable vector search, and cost-effective training. For example, media and entertainment teams can now handle high-resolution video workflows, while data teams benefit from improved analytics pipelines. The integration of NVIDIA cuVS into OpenSearch Serverless also simplifies the transition from raw data to production-ready AI systems, reducing operational overhead.
AWS’s work with NVIDIA offers a streamlined path for organizations to scale AI initiatives without increasing operational complexity.
What Are the Broader Implications for the AI Industry?
The collaboration highlights the reliance on specialized hardware and cloud partnerships to meet AI demand. By embedding GPU acceleration into core services like OpenSearch and EC2, AWS is providing enterprises more practical paths to deploy AI at production scale.
What’s Next for NVIDIA and AWS Collaboration?
AWS plans to expand support for G7 instances to Amazon SageMaker AI. Both companies emphasized ongoing co-engineering efforts to address emerging AI workloads.