Amazon Web Services (AWS) has recently announced the launch of "AI Factories," a new on-premises product developed in collaboration with Nvidia. This innovative offering allows large corporations and government entities to deploy and manage advanced AI systems directly within their own data centers, specifically addressing growing concerns around data sovereignty and control.
The core idea behind AWS AI Factories is to cater to organizations that require absolute control over their data, preventing it from residing with third-party model makers or even sharing hardware with external entities. Customers supply the necessary power and data center infrastructure, while AWS provides and manages the AI system, seamlessly integrating it with other AWS cloud services.
The product name "AI Factories" might sound familiar, and for good reason. It's a term Nvidia uses for its comprehensive hardware systems designed to run AI workloads, encompassing everything from GPU chips to networking technology. This AWS AI Factory is, in fact, a direct collaboration, combining AWS's technological prowess with Nvidia's specialized AI infrastructure, as both companies confirm.
The AWS Factory leverages a powerful combination of AWS and Nvidia technologies. Organizations deploying these systems can choose between Nvidia's latest Blackwell GPUs or Amazon's new Trainium3 chips. The solution integrates AWS's homegrown networking, storage, databases, and security features, and can tap into Amazon Bedrock for AI model selection and management, as well as AWS SageMaker AI for model building and training.
Interestingly, AWS is not the only major cloud provider investing heavily in on-premises Nvidia AI Factories. Microsoft, for instance, showcased its own deployment of Nvidia AI Factories in its global data centers in October, primarily to run OpenAI workloads. While Microsoft initially focused on internal use, it has also outlined solutions for data sovereignty, including "Azure Local," which involves Microsoft-managed hardware installed directly at customer sites. This strategy was detailed in an announcement regarding comprehensive sovereign solutions for European organizations.
This trend highlights an intriguing irony: the rise of AI is prompting the biggest cloud providers to significantly invest in corporate private data centers and hybrid cloud solutions, reminiscent of the industry landscape over a decade ago. It underscores a renewed focus on data control and localized processing for critical AI workloads.
