Amazon Web Services unloaded its annual barrage of new development tools and services this week at re:Invent 2024 in Las Vegas, where Christmas came three weeks early for some people. Developers will unwrap a wonderland of new options due to what is now becoming available from the world's largest cloud computing software provider.

In his keynote address on Day 2, AWS's first-year CEO Matt Garman introduced a sweeping array of new products, emphasizing AWS's focus on its developer community, use of AI, customer-centricity, and inventiveness.

“The way the whole AWS developer community has grown is really incredible,” Garman told a full house of technologists at The Venetian Conference Center. “We now have 600 user groups. They're in 120 countries, and the entire AWS community comprises many millions all around the world. And one of the great things about that community is that we get feedback from you that goes directly into the products and informs what we build and announce here today.”

AWS Puts Money Where Its GenAI Mouth Is Underscoring AWS's focus on democratizing access to generative AI as the defining technology of the mid-2020s, Garman announced a $1 billion investment in global startups in 2025 by AWS, providing credits to fuel their exploration and innovation in generative AI.

“They push us to innovate. They innovate on top of us and they move really fast and we learn a ton from startups,” Garman said. “Now, with generative AI, there is never a more exciting time out there in the world to be a startup. Generative AI has the potential to disrupt every industry. And when you look, disruption comes from startups.”

“So, it's a fantastic time if you're a startup to really be thinking about how you disrupt Industries.”

Key Software News Headlining the new AI assets, AWS unveiled Amazon Nova, which Garman described as “a new generation of AI foundation models that have state-of-the-art intelligence across a wide range of tasks.”

These models include Amazon Nova Micro (a fast text-to-text model) and Amazon Nova Lite, Amazon Nova Pro, and Amazon Nova Premier (multi-modal models that can process text, images, and videos to generate text). The company also launched Amazon Nova Canvas (which generates studio-quality images) and Amazon Nova Reel (which generates studio-quality videos).

Amazon Nova will be available on Amazon Bedrock, a managed service that makes high-performing foundation models from leading AI companies and Amazon available for use through a unified API.

Garman introduced Bedrock Agents, which enable the creation of intelligent agents that can execute tasks across various systems and data sources. With the new multi-agent collaboration feature, Bedrock Agents can now orchestrate complex workflows involving hundreds of agents, unlocking new possibilities for automation and intelligent decision-making.

S3 Continues to evolve S3 (Simple Storage Service) is the SaaS app upon which AWS built its business beginning in 2006, and it's still being updated 18 years later. AWS announced new S3 features that make it the first cloud object store with fully-managed support for Apache Iceberg for faster analytics, in addition to being an optional way to store and manage tabular data at any scale.

With S3 Table Buckets, AWS introduces a new bucket type specifically designed for iceberg tables, delivering 3x better query performance and 10x higher transactions per second for data lake workloads.

In addition, Garman said that S3 Metadata smoothes out data discovery in near real-time by automatically capturing queryable object metadata, and custom metadata using object tags, storing it in S3 Tables for accelerating analytics across data lakes. Developers have been asking for this for a while, he said.

Databases get fancier In the database department, new capabilities have been released for Amazon Aurora and Amazon DynamoDB to support user workloads that need to operate across multiple regions, whether they use SQL or NoSQL. Amazon Aurora DSQL is a new serverless, distributed SQL database that, according to AWS, enables customers to build applications with high availability, strong consistency, PostgreSQL compatibility, and 4X faster reads and writes compared to other distributed SQL databases.

There's a newly updated version of Amazon SageMaker, which AWS says will unify capabilities for faster SQL analytics, petabyte-scale big data processing, data exploration and integration, model development and training, and generative artificial intelligence into one integrated platform.

Hardware, too AWS started toying with the idea of making its own processors about five years ago and came up with its Inferencia and Trainium ML chip lines for training deep-learning models. The idea was to manufacture specialized processors for specialized AWS workloads, to save users time and money, Garman said.

AWS announced the general availability of Trainium 2, AWS's most powerful AI training and inference chip to date. AWS claims that Trainium 2 enables users to train and deploy cutting-edge generative AI models at significantly lower costs, Garman said.

Looking ahead, Garman said Trainium 3 is set to launch later next year, promising a 2x performance boost and 40% greater efficiency compared to Trainium 2 – which itself just came out. Progress in this sector is rapid.

This week the company also unveiled the general availability of Trainium2-powered Amazon EC2 instances and new Trn2 UltraServers, which enable users to train and deploy the newest AI models in addition to future large language models (LLM) and foundation models. For a complete list of announced new products and services from the re:Invent 2024 conference, go here.