Top takeaways from AWS re:Invent 2025

AWS re:Invent is always packed full of new product announcements and tech insights. The 2025 conference was no exception. If you couldn’t make it out to Las Vegas or just want a refresher, here are the top takeaways from the event. 

Explore the rest of our 2025 re:Invent coverage:

• AWS re:Invent 2025: Takeaways from Matt Garman’s keynote

• Pluralsight recognized as AWS Emerging Business Applications Partner of the Year

In the past, Amazon S3 objects were limited to 5 TB. Now, the maximum has been increased to 50 TB for a single object. This vastly optimizes processing and performance of large objects like AI training datasets, high-resolution videos, and more. 

Learn more about the increased Amazon S3 maximum object size.

Government agencies or organizations in highly regulated industries often need to keep sensitive data on-premises for security or compliance reasons. If they want to use AI, they can’t use public clouds. 

Now, these organizations (and any organization that wants complete control over their data), can use the new AWS AI Factories.

This new service allows organizations to run AWS AI services and infrastructure within their own data centers. Customers supply the data centers and power, and AWS provides the hardware and software, including storage, compute, and AI services like Bedrock and SageMaker AI.

Learn more about AWS AI Factories.

AWS introduced their new Trainium3 chip and teased the forthcoming Trainium4. Trainium3 powers Amazon EC2 Trn3 UltraServers for faster, lower-cost generative AI training.

It’s worth noting the origin of these chips. Trainium chips were originally designed for training models, but they’re now being used for the majority of Bedrock inference. In the future, we might also see them being used for complex analytics or high performance computing workloads.

Learn more about AWS Trainium3.

Training an AI model so it’s actually useful for your organization can be expensive and lead to catastrophic forgetting. Amazon Nova Forge is a new service that aims to solve this problem. 

Organizations can start with pre-trained, mid-trained, or post-trained models and then blend their datasets with Amazon Nova-curated training data. The result is frontier models that incorporate your organization’s specialized knowledge and proprietary data—without forgetting foundational skills.

Learn more about Nova Forge.

Will AI take developers’ jobs? Maybe. 

Will AI make developers obsolete? “Absolutely not, if you evolve,” said Dr. Werner Vogels, CTO and VP of Amazon.

In his final re:Invent closing keynote, he emphasized the need for Renaissance Developers in a world of AI. The key traits of Renaissance Developers include:

1. Curiosity: Curiosity leads to innovation and invention. Experiment, be willing to fail, and put yourself in positions that test you.

2. Think in systems, not isolated parts: Look beyond your own work to see how it impacts the overall system. That will allow you to make changes that make a difference.

3. Communication: Interpersonal communication is important. But communication is also important for building with AI. Devs now communicate with machines in natural language, and that’s much more ambiguous than code. Strong communication skills will reduce ambiguity and allow you to build better and faster.

4. Ownership: AI tools let developers build and move faster, but it’s still your responsibility to verify and take ownership of software quality. Use durability and code reviews to validate AI-generated code and catch hallucinations and inaccuracies before they’re deployed.

5. Polymath: Expand your knowledge beyond deep domain expertise. T-shaped developers often make better architectural choices because they understand how their work shapes the overall system.

Start your Renaissance Developer journey and learn AWS, cloud, AI, and software development skills with Pluralsight’s hands-on platform.

Despite the flashy AI announcements, many practitioners wished AWS had connected more of their new services back to core infrastructure and cloud fundamentals. AI and AI agents will undoubtedly play a major role in innovation in the future, but the underlying infrastructure is what will allow IT practitioners to make the most of new technology. 

If you’re feeling anxious about AI, try not to lose sight of the valuable work you do. 

“[AI doesn’t] remove the work that only you can do. Remember, the work is yours, not that of the tools. It is your work that matters,” said Vogels.

Learn more about Pluralsight’s partnership with AWS and explore Pluralsight on the AWS Marketplace.