Matt Garman keynote takeways

Martin Nanchev - Dec 3 - - Dev Community

Matt Garman unveiled at AWS re:Invent 2024 a suite of transformative updates across compute, storage, databases, AI, and real-world applications, redefining what's possible in the cloud. Here’s a deep dive into the innovations that can drive your business forward.

🚀Compute Breakthroughs

  • P6 EC2 Instances
    With NVIDIA’s most advanced GPUs, these instances deliver 2.5x faster machine learning (ML) performance at reduced costs compared to the P5.

  • Amazon EC2 Trainium 2 Instances
    Achieving 20.8 petaflops, these ML powerhouses offer 30–40% better performance, enabling partners like Databricks to accelerate workloads.

https://aws.amazon.com/about-aws/whats-new/2024/12/amazon-ec2-trn2-instances-available/

  • Trainium Ultra Server
    Featuring 83 petaflops, this single-node ultra server supports high-performance computing for ML and LLM models at unprecedented speed.

  • Trainium 3 (Coming in 2025)
    Built on a 3nm process, it’s set to deliver 40% greater efficiency, setting a new benchmark for ML hardware.

💾 Storage Revolution

  • Amazon S3 Table Buckets Turn S3 into a high-performance data lake! With Apache Iceberg, you can run SQL queries directly, achieving 3x faster queries and 10x higher transactions per second.

https://aws.amazon.com/about-aws/whats-new/2024/12/amazon-s3-tables-apache-iceberg-tables-analytics-workloads/
https://aws.amazon.com/blogs/aws/new-amazon-s3-tables-storage-optimized-for-analytics-workloads/

  • S3 Metadata Discovery Simplify object discovery with SQL-accessible metadata stored in Iceberg tables, unlocking faster, more intuitive workflows.

https://aws.amazon.com/about-aws/whats-new/2024/12/amazon-s3-metadata-preview

  • Zero-ETL integration for applications stored across all SAAS applications, aurora, redshift, dynamodb

📊 Database Innovations

  • Amazon Aurora DSQL
    A multi-region, low-latency database offering 5-nines availability. Aurora DSQL uses satellite time sync to nearly break the CAP theorem and is 4x faster than Google Spanner.
    https://aws.amazon.com/about-aws/whats-new/2024/12/amazon-aurora-dsql-preview/

  • DynamoDB Global Tables
    These active-active, highly available tables ensure strong consistency and resilience for mission-critical workloads.

https://aws.amazon.com/about-aws/whats-new/2024/12/amazon-dynamodb-global-tables-previews-multi-region-strong-consistency

🧠 AI & Inference Upgrades

🏆 Real Stories of Impact

Apple & AWS Collaboration
From Graviton 3 optimizations to accident detection using K-nearest neighbors, AWS drives ML innovation for Apple’s private LLMs, hosted on Trainium instances.

JP Morgan Chase
Transitioning 6,000 applications to AWS, the financial giant enhances resilience, reduces costs, and supports global business growth.

Genentech
Transforming drug discovery with generative AI, their systems enable scientists to ask complex questions and receive actionable insights in real-time.

PagerDuty Advance - the new GenAI PagerDuty, that keep you well informed and help you follow and engage issues. Less time for issues and more time for building with Amazon Q for PagerDuty Advance

The Road Ahead
AWS continues to redefine the limits of cloud innovation. Whether you’re modernizing infrastructure, exploring advanced AI models, or optimizing workloads, these advancements empower organizations to scale faster, innovate deeper, and inspire change.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .