Table of Contents
According to CEO Matt Garman, AWS is focused on helping organizations “run their largest and most demanding applications”
In sum – what to know:
Enterprise AI adoption is accelerating – Garman said customers are moving from experimentation to production, demanding easier integration of generative AI into business workflows.
AWS expands managed and automated AI services – More tools aim to simplify deployment, tuning, and connection to customer data.
Broader model and accelerator support remains central – Garman highlighted AWS’s 15-year collaboration with Nvidia and multi-chip strategy to meet diverse AI workload needs.
Amazon Web Services (AWS) used its re:Invent 2025 keynote to outline how the company is reshaping its AI software stack and managed services to help enterprises deploy models more quickly and integrate generative AI into existing applications.
While CEO Matt Garman devoted part of his remarks to long-term infrastructure planning, he also focused on how customers are adopting AI across industries and what AWS is building to support that shift.
Garman said enterprises are increasingly incorporating generative AI into core business workflows, a sign that the technology is moving beyond experimentation and into production. He noted that customers want AI systems that are easier to deploy, tune, and connect to their own data, and emphasized that AWS is working to simplify that process. “Customers want more accessibility,” he said, pointing to growing demand for tools that reduce the operational complexity of running large-scale models.
AWS’s approach includes expanding its managed AI services and building more automation into model deployment. Garman said the company continues to work closely with major model developers, including long-standing partners such as Nvidia. “We’ve been working closely with Nvidia for more than 15 years,” he said, adding that AWS will keep offering multiple accelerators and model options to support different workloads.
He also linked the evolution of these tools to broader customer expectations around performance and reliability. According to Garman, AWS is focused on helping organizations “run their largest and most demanding applications” while offering a software environment that allows them to scale without redesigning their systems. That includes enhancing developer capabilities and expanding foundational services that sit beneath AI workloads, enabling faster integration of models into enterprise applications.
Garman added that as model adoption accelerates, customers increasingly want predictable access to compute and streamlined ways to build AI-powered features into their products. AWS’s goal, he said, is to give organizations flexible pathways to deploy generative AI at their own pace while ensuring the underlying platform can support long-term growth.
During the Re:Invent 2025 conference in Las Vegas, AWS has also introduced AWS AI Factories, a new offering that delivers AI-ready infrastructure directly into customers’ own data centers.
The AI Factory model allows organizations to deploy AWS-managed AI hardware and services on-premises. The setup includes Nvidia GPUs, AWS Trainium processors, and the company’s networking, storage, and database technologies.
The dedicated environments are operated solely for each customer, giving governments and large enterprises the ability to scale AI workloads while meeting strict compliance, data-sovereignty, and security requirements.
Designed to function similarly to a private AWS Region, the AI Factories offering provide access to AWS managed services — including foundation models — while allowing customers to determine exactly where data is processed and stored.
During his keynote, Garman also said the industry is entering a new phase defined by “a major computing shift driven by accelerated AI,” requiring sustained global capacity expansion.
