AWS kicked off re:Invent 2025 in Las Vegas on Tuesday (Dec. 2) with high expectations. As the artificial intelligence (AI) race heats up, Amazon needs to prove that its years of heavy investment in model infrastructure and automated systems are paying off. The company is going up against Microsoft and Google, both of which are rapidly expanding their cloud AI services.
This annual showcase is crucial because AWS contributes significantly to Amazon’s revenue and a greater share of its profits, making the success of its AI strategy vital to the company’s future.
AWS CEO Matt Garman, in his keynote, highlighted the scale of the business, stating, “AWS has grown to be a $132 billion business, accelerating by 20% year over year. The amount we grew in the last year alone is larger than the annual revenue of more than half of the Fortune 500.”
The announcements on Day 1 demonstrated how AWS is reshaping its infrastructure for the next phase of cloud and AI growth.
Agentic AI stood out as a key theme of the session. AWS made significant updates to Amazon Bedrock and introduced its new Nova models, calling them essential for creating autonomous enterprise systems. These updates support AI agents that can interpret signals, plan steps and complete tasks across supply chains, engineering and customer support.
Garman tied these advances to AWS’ engineering philosophy, explaining, “There are no shortcuts in AI infrastructure. You have to optimize every single layer of hardware and software, and that is something only AWS does.”
AWS also introduced AI Factories, a new option that brings dedicated AWS AI infrastructure straight to an enterprise or government data center. The model is designed for organizations that need to scale AI while following strict sovereignty and compliance rules. By functioning like a private AWS Region with local computing, storage, databases and AI services, AI Factories provide customers secure, low-latency access to AWS capabilities without transferring sensitive data offsite.
Customer use cases showed how businesses are using this expanded architecture. Sony is scaling its enterprise AI systems on AWS to unify data sources, automate production processes, and improve engagement across global divisions, according to the update on Sony’s enterprise AI platforms. Adobe also strengthened its partnership with AWS to scale Firefly models for creative and marketing tasks.
AWS also unveiled new tools to improve operational resilience and support outage recovery. The company announced a cloud AI system that helps engineers recover from outages by analyzing event patterns, detecting configuration changes, and recommending steps to fix issues. As CNBC reported, the system eases engineers’ burden during major incidents and reflects lessons learned from managing one of the world’s largest distributed platforms.
Garman noted that customers are still in the early stages of seeing measurable benefits from generative AI. Still, he added, “When I speak to customers, you have not yet seen returns that match the promise of AI. The true value has not yet been unlocked, but that is changing quickly.”
What Else Stood Out
- AWS showcased its new Trainium 3 chips, which it described as faster, more efficient AI training processors. The company claimed these chips help lower the time and cost of developing large models, a significant challenge for businesses investing in generative AI.
- AWS introduced new UltraServer systems built around these chips. These servers are designed to provide companies with a simpler way to run demanding AI projects without needing to build or manage advanced hardware themselves.
- The company announced Nova Forge, a new tool that assists businesses in customizing AI models with their own data. AWS stated this service is for organizations looking for tailored AI results without requiring internal research teams.
- AWS expanded AgentCore’s capabilities within Amazon Bedrock. The updates aim to help companies create AI agents that can perform more complex tasks like document analysis, task planning and multistep workflows.
- AWS also announced new modernization tools that help organizations transition older systems to cloud environments that can support modern AI workloads. These updates are designed to simplify the process for companies still relying on legacy software.
Source: https://www.pymnts.com/
