Top stories from AWS’s major technology showcase re:Invent 2025

Amazon Web Services’ annual AWS re:Invent technical conference wrapped up another day with a deluge of product news and key insights — plus the obligatory customer success stories.

The unsurprising topic is enterprise AI. This year it’s all about upgrades that give customers more control to customize AI agents, including one that AWS claims can learn from you and then work autonomously for days.

AWS re:Invent 2025, which runs through December 5, kicked off with a keynote from AWS CEO Matt Jarman, who weighed in on the idea that AI customers can unlock the “true value” of AI.

“AI assistants are starting to give way to AI agents that can perform tasks and automate on your behalf,” he said during his Dec. 2 keynote. “This is where we start to see material business returns from your AI investments.”

On December 3, the conference continued messaging AI agents as well as delving into customer stories. One of the keynote talks was given by Swami Sivasubramanian, Vice President of Agentic AI at AWS. To say he was optimistic might be an understatement of the prevailing atmosphere.

“We are living in a time of great change,” Sivasubramanian said during the talk. “For the first time in history, we can describe what we want to achieve in natural language, and agents create the plan. They write the code, call in the necessary tools and implement the complete solution. Agents give you the freedom to build without limits, accelerating how quickly you can go from idea to impact in a big way.”

While AI Agent news promises a continued presence during AWS re:Invent 2025, there have been other announcements as well. Here’s a roundup of the ones that caught our attention. TechCrunch will update this article with the latest insights at the top, through the end of AWS re:Invent. Make sure to check back.

Double the level of LLMs

AWS announced more tools for enterprise customers to create their own models. Specifically, AWS said it is adding new capabilities for both Amazon Bedrock and Amazon SageMaker AI to make it easier to create custom LLM courses.

For example, AWS brings serverless model customization to SageMaker, allowing developers to start building a model without having to think about compute resources or infrastructure. Serverless model customization can be accessed either through a self-directed path or by prompting an AI agent.

AWS also announced Reinforcement Fine Tuning in Bedrock that allows developers to choose a pre-defined workflow or bounty system and have Bedrock automate their customization process from start to finish.

Andy Jassy shares some numbers

Amazon CEO Andy Jassy took to social media platform X to explain the keynote delivered by AWS President Matt Jarman. The message: The current generation of Nvidia’s competing Trainium2 AI chip is already bringing in a lot of money.

His comments were related to the unveiling of the next-generation Trainium3 chip and were intended to predict a promising revenue future for the product.

Access database savings

Among the dozens of ads, there’s one item that’s actually getting a thumbs up. Discounts.

Specifically, AWS said it will launch database savings plans, which help customers reduce database costs by up to 35% when they commit to a fixed amount of usage ($/hour) over the course of one year. The savings will be automatically applied hourly to eligible usage across supported database services, and any additional usage beyond the commitment is billed at on-demand rates, the company said.

Corey Quinn, chief cloud economist at Duckbill, summed it up well in his blog post titled “Six Years of Complaining Finally Pays Off.”

You can’t get a better deal than free, Amazon hopes

Is there any way for another AI-based programming tool to win the hearts of startup founders? Amazon is hoping a year of free credits will solve the problem with its offering, Kiro. The company will award Kiro Pro+ credits to eligible startups that apply for the deal before the end of the month. However, only early-stage startups in some countries are eligible.

AI training chip with Nvidia compatibility

AWS has introduced a new version of its AI training chip called Trainium3 along with the AI ​​system called UltraServer that powers it. TL;DR: This upgraded chip comes with some impressive specs, including the promise of up to 4x performance gains for both AI training and inference while reducing power usage by 40%.

AWS also made a teaser. The company already has Trainium4 software in development, which will be able to work with Nvidia chipsets.

Expanding AgentCore capabilities

AWS has announced new features in its AgentCore AI agent building platform. One noteworthy feature is Policy in AgentCore, which gives developers the ability to set limits for AI agents more easily.

AWS also announced that agents will now be able to record and remember things about their users. In addition, it announced that it will help its clients evaluate agents through 13 pre-set rating systems.

A non-stop AI worker bee

AWS announced three new AI agents (there’s that term again) called “Frontier Agents,” including one called the “Kiro Autonomous Agent” that writes code and is designed to learn how the team likes to work so it can work largely on its own for hours or days.

One of these new agents handles security operations like code reviews, and a third handles DevOps tasks like preventing incidents when new code is live. Preview versions for agents are available now.

New Nova models and services

AWS is introducing four new AI models within the Nova AI family of models – three dedicated to text generation and one that can generate text and images.

The company also announced a new service called Nova Forge that allows AWS cloud customers to access pre-trained, mid-trained, or post-trained models that they can then improve by training on their proprietary data. AWS’s big offering is flexibility and customization.

Lyft’s argument for AI agents

The ride-hailing company was among several AWS customers who descended on the event to share their success stories and evidence of how the products have impacted their businesses. Lyft uses Anthropic’s Claude model via Amazon Bedrock to create an AI agent that handles driver and passenger questions and issues.

The company said this AI agent reduced the average resolution time by 87%. Lyft also said it saw a 70% increase in driver use of its AI agent this year.

AI factory for private data center

Amazon also announced “AI factories” that allow major companies and governments to run AWS AI systems in their own data centers.

The system was designed in partnership with Nvidia and includes both Nvidia and AWS technology. While companies using it can stock it with Nvidia GPUs, they can also opt for Amazon’s latest home-made AI chip, the Trainium3. The system is Amazon’s way of addressing data sovereignty, or the need for governments and many companies to control their data and not share it, even for the use of artificial intelligence.

See the latest reveals on everything from agent AI and cloud infrastructure to security and more from Amazon Web Services’ flagship event in Las Vegas. This video is brought to you in partnership with AWS.