Amazon Web Services (AWS) has announced a sweeping suite of new AI products and infrastructure advancements, headlined by the groundbreaking Amazon Nova family of foundation models, promising to redefine the AI landscape with unprecedented price-performance.
AWS recently held a media roundtable in Makati City, Philippines to make the announcements.
At the core of the announcement is Amazon Nova, a collection of state-of-the-art foundation models designed to deliver “frontier intelligence” at significantly reduced costs. Available through Amazon Bedrock, Nova models cater to diverse needs, from rapid text generation with Amazon Nova Micro to complex multimodal reasoning with Amazon Nova Premier.
Paul Chen, Head of Solutions Architects, ASEAN, AWS shared: “Amazon Nova models, particularly Micro, Lite, and Pro, are at least 75 percent less expensive than comparable models in Bedrock, while also being the fastest in their respective intelligence classes.”
The Nova family also introduces specialized models like Amazon Nova Canvas for studio-quality image generation and Amazon Nova Reel for high-fidelity video creation.
Bedrock platform
AWS is also strengthening its Amazon Bedrock platform with industry-first AI safeguards, agent orchestration capabilities, and advanced customization options. The platform now boasts an expanded selection of leading models, including Luma AI’s Ray 2, poolside’s malibu and point, and Stability AI’s Stable Diffusion 3.5 Large.
Furthermore, the launch of the Amazon Bedrock Marketplace brings in diverse models, including Asia-Pacific region providers like Upstage’s Solar Pro for translation and AI Singapore’s SEA-LION for Southeast Asian languages.
Transforming workloads with Generative AI
AWS is leveraging generative AI to streamline complex workloads. Amazon Q Developer is now enhanced to simplify application migration and modernization, saving significant time and resources. Amazon Q Business, Amazon Q in QuickSight, and integrations with third-party applications like Asana and Zoom now share a unified enterprise data index. Over 50 new actions are available to automate tasks across various applications.
Next-generation infrastructure and Silicon for AI
To support the demands of advanced AI, AWS has announced the general availability of AWS Trainium2-powered Amazon EC2 instances and unveiled the next-generation Trainium3 chips. New Trn2 UltraServers are designed for training and deploying the latest LLMs and FMs with exceptional performance and cost efficiency.
AWS has also introduced cutting-edge data center innovations, including liquid cooling for high-density AI servers, optimized power utilization, and advanced control systems, ensuring 99.9999% infrastructure availability. These improvements also reduce the potential impact of electrical issues by 89%.
Data as the backbone of AI innovation
AWS has unveiled the next generation of Amazon SageMaker, a unified platform for data, analytics, and AI. The new SageMaker Unified Studio simplifies data access and integration, while SageMaker Catalog ensures secure data and model discovery. SageMaker Lakehouse provides unified access to data across various storage solutions.
Enhanced Amazon SageMaker HyperPod capabilities are designed to reduce model training time and cut costs by up to 40%. AWS has also enhanced Amazon DynamoDB, Amazon Aurora, and Amazon S3 to support demanding workloads, with S3 now offering fully-managed support for Apache Iceberg.
These announcements solidify AWS’s commitment to driving AI innovation by providing powerful, cost-effective, and accessible tools for developers and businesses alike.
READ MORE TECH NEWS.
