Expanding what’s possible for developers and businesses in the cloud, NVIDIA and Amazon Web Services are converging this week at AWS re:Invent in Las Vegas to showcase new solutions designed to accelerate breakthroughs in AI and robotics and simplify research in the development of quantum computing.
AWS re:Invent is a conference for the global cloud computing community, packed with keynotes and over 2,000 technical sessions.
Highlights of the announcement include the availability of NVIDIA DGX-Cloud on AWS and enhanced AI, quantum computing and robotics tools.
NVIDIA DGX Cloud on AWS for AI at Scale
THE NVIDIA DGX-Cloud The AI Computing Platform is now available through AWS Marketplace Private Offerings, providing a high-performance, fully managed solution for businesses to train and customize AI models.
DGX Cloud offers flexible terms, a fully managed and optimized platform, and direct access to NVIDIA experts to help businesses rapidly scale their AI capabilities.
Leonardo.ai, an early adopter of the Canva family, is already using DGX Cloud on AWS to develop advanced design tools.
AWS Liquid-Cooled Data Centers with NVIDIA Blackwell
Newer AI servers benefit from liquid cooling to more efficiently cool high-density computing chips for improved performance and power efficiency. AWS has developed solutions that provide configurable liquid-to-chip cooling in its data centers.
The cooling solution announced today will seamlessly integrate air and liquid cooling capabilities for the most powerful rack-scale AI supercomputing systems, such as NVIDIA GB200 NVL72as well as network switches and storage servers from AWS.
This flexible, multi-modal cooling design provides maximum performance and efficiency for running AI models and will be used for the next generation. NVIDIA Blackwell platform.
Blackwell will be the basis for Amazon EC2 P6, DGX Cloud instances on AWS and Ceiba Project.
NVIDIA Advances Physical AI with Accelerated Robot Simulation on AWS
NVIDIA is also expanding the reach of NVIDIA Omniverse on AWS with NVIDIA Isaac Simnow running on high-performance Amazon EC2 G6e instances accelerated by NVIDIA L40S GPU.
Available now, this reference application is based on NVIDIA Omniverse allows developers to simulate and test AI-driven robots in physical virtual environments.
One of the many workflows enabled by Isaac Sim is synthetic data generation. This pipeline is now further accelerated with the infusion of OpenUSD NIM microservices, from scene creation to data augmentation.
Robotics companies such as EscapeCohesive Robotics, Cobot, Field AI, Standard Bots, Swiss Mile and Vention use Isaac Sim to simulate and validate the performance of their robots before deployment.
Additionally, Rendered.ai, Software server and Tata Consultancy Services use the synthetic data generation capabilities of Omniverse Replicator and Isaac Sim to bootstrap perceptual AI models that power various robotics applications.
NVIDIA BioNeMo on AWS for advanced AI-driven drug discovery
NVIDIA BioNeMo NIM Microservices and AI Blueprintsdeveloped to advance drug discovery, are now integrated with AWS HealthOmicsa fully managed biological data computing and storage service designed to accelerate scientific advances in clinical diagnostics and drug discovery.
This collaboration provides researchers with access to AI models and scalable cloud infrastructure suitable for drug discovery workflows. Several biotechnology companies are already using NVIDIA BioNeMo on AWS to drive their research and development pipelines.
For example, Bio A-Alphaa Seattle-based biotechnology company, recently released a study at biorxiv describing a collaborative effort with NVIDIA and AWS to develop and deploy an antibody AI model called AlphaBind.
Using AlphaBind via the BioNeMo framework on Amazon EC2 P5 instances equipped with NVIDIA H100 Tensor Core GPUA-Alpha Bio achieved 12x inference speed and processed over 108 million inferences calls in two months.
In addition, Software server today launched Drug Discovery, its generative AI solution built with NVIDIA Blueprints, to enable computer-aided drug discovery and efficient drug development. This solution is expected to provide faster workflows and will soon be available on AWS Marketplace.
Real-time AI plans: Ready-to-deploy options for video, cybersecurity, and more
NVIDIA’s latest AI models are available for instant deployment on AWS, creating real-time applications such as vulnerability scanning for container securityand easily accessible video search and synthesis agents.
Developers can easily integrate these plans into existing workflows to speed up deployments.
Developers and businesses can use the NVIDIA AI Blueprint for Video Search and Summarization to create visual AI agents that can analyze real-time or archived video to answer user questions, generate summaries, and activate alerts for specific scenarios.
AWS collaborated with NVIDIA to provide a reference architecture applying the NVIDIA AI Blueprint for vulnerability scanning to augment early security patches in continuous integration pipelines across AWS native cloud services.
NVIDIA CUDA-Q on Amazon Braket: Quantum Computing Made Practical
NVIDIA CUDA-Q is now integrated with Amazon Braket to streamline the development of quantum computing. CUDA-Q users can use Amazon Braket’s quantum processors, while Braket users can leverage CUDA-Q’s GPU-accelerated workflows for development and simulation.
The CUDA-Q platform allows developers to create hybrid quantum-classical applications and run them on many different types of quantum, simulated, and physical processors.
Now preinstalled on Amazon SupportCUDA-Q provides a transparent development platform for hybrid quantum-classical applications, paving the way for new potentials in quantum research.
Enterprise Platform Providers and Consulting Leaders Advance AI with NVIDIA on AWS
The world’s leading software platforms and systems integrators help businesses rapidly scale generative AI applications built with NVIDIA AI on AWS to drive innovation across industries.
Cloudera uses NVIDIA AI on AWS to enhance its new AI inference solution, helping Mercy Corps improve the accuracy and efficiency of its aid delivery technology.
Cohesion has integrated NVIDIA NeMo Retriever microservices into its AI-powered generative conversational search assistant, Cohesity Gaia, to improve retrieval-augmented recall performance. Cohesity customers running on AWS can take advantage of NeMo Retriever integration in Gaia.
DataStax announced that Wikimedia Deutschland is applying the DataStax AI platform to make Wikidata available to developers as an integrated vectorized database. The Datastax AI platform is built with NVIDIA NeMo Retriever and NIM microservices, and available on AWS.
Deloitte AI C-Suite now supports NVIDIA AI Enterprise software, including NVIDIA NIM microservices and NVIDIA NeMo for use cases specific to CFOs, including financial statement analysis, scenario modeling, and market analysis.
RAPIDS Quick Start Notebooks are now available on Amazon EMR
NVIDIA and AWS also accelerate data science and data analytics workloads with RAPIDS Accelerator for Apache Spark, which Accelerates analytics and machine learning workloads without code changes and reduces data processing costs by up to 80%.
Quick Start Notebooks for RAPIDS Accelerator for Apache Spark are now available on Amazon DMEAmazon EC2 and Amazon EMR on EKS. These provide a simple way to qualify Spark jobs optimized to maximize RAPIDS performance on GPUs, all within AWS EMR.
NVIDIA and AWS Power the Next Generation of Cutting-Edge Industrial Systems
THE NVIDIA IGX Orin And Jetson Orin platforms now integrate seamlessly with AWS IoT Green Grass to streamline the deployment and execution of AI models at the edge and efficiently manage fleets of connected devices at scale. This combination improves scalability and simplifies the deployment process for industrial and robotics applications.
Developers can now leverage NVIDIA’s advanced computing power with AWS’s purpose-built IoT services, creating a secure, scalable environment for autonomous machines and intelligent sensors. A getting started guide, written by AWS, is now available to help developers implement these features.
This integration highlights NVIDIA’s work to advance edge-ready industrial systems to enable fast, intelligent operations in real-world applications.
NVIDIA NIM Microservices Now Available on Amazon Bedrock Marketplace
Finally, on December 4, AWS oneannounced that NVIDIA NIM Microservices are now available directly from the Amazon Bedrock Marketplace, providing developers with simplified access to inference powered by NVIDIA. This extension supports faster AI inference and lower latency for AI agents, AI physics applications, and more. NVIDIA NIM is part of the NVIDIA AI Enterprise software platform, available through the AWS Marketplace.
This update simplifies the deployment of NIM microservices, including NVIDIA Nemotron-4, at scale. These offerings are also accessible through AWS Marketplace and SageMaker Jumpstart.
Learn more about NVIDIA’s work on AWS: re:Invent 2024 through live demos, technical sessions, and hands-on workshops.
Editor’s Note: This blog has been updated to reflect the announcement on Wednesday, December 4, that AWS will extend NVIDIA NIM microservices to core AWS AI services.
See notice regarding software product information.