NVIDIA and AWS Collaborate to Bring AI to Production at Scale

NVIDIA and Amazon Web Services are expanding their collaboration with new EC2 G7 instances and GPU-accelerated vector search in OpenSearch Serverless, aiming to streamline production-scale AI deployment for enterprise customers [1]. The new Amazon EC2 G7 instances are powered by NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs, targeting AI inference, graphics, video, and data analytics workloads [1]. Compared with the previous-generation G6 instances, the G7 delivers up to 4.6x AI inference performance and up to 2.1x graphics performance, according to NVIDIA [1]. The instances support configurations with up to eight GPUs, 256GB of total GPU memory, 700 Gbps of EFA-enabled networking, and up to 7.6TB of local NVMe SSD storage [1]. NVIDIA, headquartered in Santa Clara, California, originally focused on GPUs for video games before broadening into AI and high-performance computing [2]. The company now controls more than 80% of the market for GPUs used in training and deploying AI models [2]. On the retrieval side, the NVIDIA cuVS library is now the default compute choice for vector indexing in Amazon OpenSearch Serverless [1]. This integration makes GPU-powered vector search a standard capability rather than a specialized optimization project [1]. NVIDIA states that vector indexing is up to 10x faster and costs a quarter of the price compared with CPU-only builds [1]. The shift enables teams building retrieval-augmented generation, semantic search, and agentic AI applications to construct billion-scale vector databases in under an hour [1]. AWS has also achieved NVIDIA Exemplar Cloud status for the NVIDIA GB300 on training workloads, signaling that the cloud provider meets NVIDIA's performance benchmarks for large-scale AI training [1]. The designation stems from co-engineering efforts between the two companies and is intended to help developers evaluate cloud infrastructure with greater confidence [1]. AWS has invested heavily in custom silicon for its cloud, including the Arm-based Graviton processor family first announced in 2018, though the latest AI-focused instances rely on NVIDIA GPUs [5]. The broader AI infrastructure market has seen rapid growth, with NVIDIA becoming the first company to surpass $5 trillion in market capitalization in 2025, driven by demand for AI data center hardware [2].