## Overview AWS re:Invent 2025 represents a strategic shift toward enterprise-grade AI implementation. The company introduced new foundation models, custom training capabilities, and next-generation infrastructure designed to give organizations full control over their AI stacks. ## What's New at re:Invent 2025 ### AI Models & Custom Training AWS launched **Amazon Nova 2**, offering improved reasoning, multimodal capabilities and cost-performance balance. The company also introduced **Amazon Nova Forge**, enabling organizations to build custom model versions using their proprietary data while controlling training phases and safety guardrails. This approach addresses enterprise needs for domain-specific AI while maintaining compliance and institutional knowledge preservation. ### Hardware & Infrastructure **Trainium3** and **Amazon EC2 Trn3 UltraServers** deliver approximately 4.4x improved compute performance over previous generations, with enhanced energy efficiency and memory/bandwidth capabilities. AWS positions custom silicon and infrastructure scaling as essential to cost-effective AI operations. ### AI + Cloud Services Ecosystem AWS expanded its AI-ready services, including enhancements to data analytics, vector databases, and cloud-native tools. This creates an end-to-end stack from custom model training through deployment. ## Key Takeaways The announcements enable "AI on enterprise terms," with emphasis on data control, model customization, and infrastructure ownership. This matters particularly for regulated industries like healthcare and education. The combination of Nova Forge, Trn3 infrastructure, and cloud services reduces friction for scaling AI production workloads across organizations.