AWS Re:Invent: Gen AI, Storage, Compute Launches
Gen AI was front and center at AWS re:Invent, including during CEO Adam Selipsky's keynote.
![Rocking out at AWS re:Invent 2023 Rocking out at AWS re:Invent 2023](https://eu-images.contentstack.com/v3/assets/blt10e444bce2d36aa8/bltfa77d0823e64a11d/65667532ad5fad040a67d797/Guitar_player_at_AWS_reInvent_2023.jpg?width=700&auto=webp&quality=80&disable=upscale)
AWS re:Invent got off to a rocking start on Nov. 28, 2023.
AWS is teaming with Nvidia on “the most advanced infrastructure for generative AI workloads with GPUs.”
The companies will bring together their technologies for training foundation models and building generative AI applications. These include Nvidia’s newest multi-node systems featuring next-generation GPUs, CPUs and AI software, to AWS Nitro System virtualization and security, Elastic Fabric Adapter (EFA) interconnect and UltraCluster scalability.
AWS will be the first cloud provider to bring NvidiaGH200 Grace Hopper Superchips with new multi-node NVLink technology to the cloud.
Additionally, Nvidia and AWS will collaborate to host Nvidia DGX Cloud, Nvidia’s AI training as a service, on AWS. It will be the first DGX Cloud featuring GH200 NVL32, providing developers the largest shared memory in a single instance.
The two are also teaming on what Nvidia calls Project Ceiba, to design the world’s fastest GPU-powered AI supercomputer.
“We are incredibly excited to build the largest AI factory Nvidia has ever built,” said Huang.
Project Ceiba is an at-scale system with GH200 NVL32 and Amazon EFA interconnect, hosted by AWS for Nvidia’s own research and development team. This supercomputer features 16,384 Nvidia GH200 Superchips and capable of processing 65 exaflops of AI.
Finally, AWS will also introduce three additional Amazon EC2 instances. P5e instances are powered by Nvidia H200 Tensor Core GPUs, for large-scale and cutting-edge generative AI and HPC workloads. G6 and G6e instances are powered by Nvidia L4 GPUs and Nvidia L40S GPUs, respectively, for applications such as AI fine tuning, inference, graphics, and video workloads.
Amazon Q is a new type of Gen AI-powered assistant designed specifically for work.
AWS said with Amazon Q, customers can get fast, relevant answers to pressing questions, generate content and take actions. These are all informed by their information repositories, code and enterprise systems.
Amazon Q can personalize its interactions to each individual user based on an organization’s existing identities, roles and permissions. Additionally, Amazon Q doesn’t use business customers’ content to train its underlying models. The idea is to bring gen AI-powered assistance to customers building on AWS, working internally, and using AWS applications for business intelligence (BI), contact centers and supply chain management.
Amazon Q is available to customers in preview, with Amazon Q in Connect generally available and Amazon Q in AWS Supply Chain coming soon.
AWS announced the general availability of what Selipsky described as “a new purpose-built S3 storage class.”
AWS is pitching Amazon Simple Storage Service (Amazon S3) Express One Zone as the lowest-latency cloud object storage available. It has data access speed up to 10 times faster and request costs up to 50% lower than Amazon S3 Standard, from any AWS Availability Zone within an AWS Region.
For the most demanding workloads, customers can choose to collocate their Amazon S3 Express One Zone data in the same AWS Availability Zone as their compute resources. This optimizes compute performance and costs while increasing data processing speed, said AWS.
Amazon S3 Express One Zone is the most performant storage class for request-intensive operations such as machine learning (ML) training and inference, interactive analytics, and media content creation.
AWS announced the next generation of two AWS-designed chip families — AWS Graviton4 and AWS Trainium2. These, it said, improve price performance and energy efficiency for a range of customer workloads, including machine learning (ML) training and gen AI applications.
Selipsky said Graviton4 provides up to 30% better compute performance, 50% more cores and 75% more memory bandwidth than current generation Graviton3 processors. They deliver the best price performance and energy efficiency for a broad range of workloads running on Amazon EC2.
Trainium2 is designed to deliver up to 4x faster training than first generation Trainium chips can be deployed in EC2 UltraClusters of up to 100,000 chips. This makes it possible to train foundation models (FMs) and large language models (LLMs) in a fraction of the time, while improving energy efficiency up to 2x.
Selipsky also unveiled new gen AI capabilities in its cloud contact center, Amazon Connect.
Powered by LLMs and other FMs available through Amazon Bedrock, the latest enhancements expand on the service’s existing ML features.
Amazon Q in Connect is a gen AI-enhanced evolution of Amazon Connect Wisdom. It provides agents with recommended responses and actions based on real-time customer questions.
Amazon Connect Contact Lens is an Amazon Connect feature that provides real-time contact center analytics and quality management. It now helps identify the essential parts of call center conversations with AI generated summaries that detect sentiment, trends, and policy compliance.
Also powered by gen AI, Amazon Lex in Amazon Connect enables contact center admins to create new chatbots and interactive voice response (IVR) systems by using natural language prompts, and improve existing systems by generating responses to commonly asked questions.
Lastly, Amazon Connect Customer Profiles is an Amazon Connect feature that enables agents to deliver faster, more personalized customer service. It can now create unified customer profiles from disparate software-as-a-service applications and databases.
AWS previewed four new capabilities for AWS Supply Chain, set for 2024.
AWS said Supply Chain Supply Planning will help customers “forecast, plan, position and replenish components and finished goods” to reduce inventory costs and respond more quickly to demand variations and supply disruptions.
AWS Supply Chain N-Tier Visibility is designed to streamline communication between customers and multiple tiers of suppliers. This is meant to improve their ability to respond to supply plans and manage demand or supply changes during the execution window. AWS said with this capability, customers can collaborate securely with their trading partners in just a few clicks.
AWS Supply Chain Sustainability will provide a central repository, enabling customers to request, collect and audit sustainability data.
Finally, Amazon Q in AWS Supply Chain provides supply chain professionals with a generative AI assistant to provide a summarized view of key risks around inventory levels, demand variability, and visualizes the trade-offs between different possible scenarios.
AWS also unveiled new integrations that enable customers to connect and analyze data without building and managing complex extract, transform, and load (ETL) data pipelines.
New Amazon Aurora PostgreSQL, Amazon DynamoDB and Amazon Relational Database Service (Amazon RDS) for MySQL integrations with Amazon Redshift make it easier to connect and analyze transactional data from multiple relational and non-relational databases in Amazon Redshift.
Customers can also now use Amazon OpenSearch Service to perform full-text and vector search on DynamoDB data in near real time. Selipsky said zero-ETL integrations help customers leverage AWS’ database and analytics services to make better data-driven decisions.
AWS also unveiled new integrations that enable customers to connect and analyze data without building and managing complex extract, transform, and load (ETL) data pipelines.
New Amazon Aurora PostgreSQL, Amazon DynamoDB and Amazon Relational Database Service (Amazon RDS) for MySQL integrations with Amazon Redshift make it easier to connect and analyze transactional data from multiple relational and non-relational databases in Amazon Redshift.
Customers can also now use Amazon OpenSearch Service to perform full-text and vector search on DynamoDB data in near real time. Selipsky said zero-ETL integrations help customers leverage AWS’ database and analytics services to make better data-driven decisions.
AWS RE:INVENT — AWS kicked off day one of AWS re:Invent Tuesday with CEO Adam Selipsky announcing a raft of new offerings, with many geared around generative AI (gen AI).
The launches include Amazon Q, a new type of Gen AI-powered assistant for work. Elsewhere, Selipsky said AWS was “reinventing storage” with the new Amazon S3 Express One Zone.
Additionally, the CEO welcomed Nvidia CEO Jensen Huang on stage to announce new joint offerings around supercomputing infrastructure, software and services for gen AI.
Elsewhere at the event, AWS channel leader Ruba Borno said the hyperscaler wants to help partners pinpoint the right gen AI solutions for their customers.
See the slideshow above to see the headlines from day one of AWS re:Invent.
About the Author(s)
You May Also Like