Amazon challenges competitors with on-premises Nvidia ‘AI Factories’

Amazon announced a new product Tuesday called “AI Factories” that allows large companies and governments to run their AI systems in their own data centers. Or as AWS puts it: customers provide the power and the data center, and AWS steps into the AI ​​system, manages it and can connect it to other AWS cloud services.

The idea is to accommodate companies and governments that are concerned with data sovereignty, or absolute control over their data, so that it cannot end up in the hands of a competitor or foreign adversary. An on-prem AI Factory means they don’t send their data to a model maker or even share the hardware.

If that product name sounds familiar, it should. That’s what Nvidia calls its hardware systems, which are packed with the tools needed to run AI, from the GPU chips to the networking technology. This AWS AI Factory is in fact a collaboration with Nvidia, say both companies.

In this case, the AWS Factory will use a combination of AWS and Nvidia technology. Companies deploying these systems can choose Nvidia’s latest Blackwell GPUs or Amazon’s new Trainium3 chip. It uses AWS’s native networking, storage, databases and security and can leverage Amazon Bedrock – the AI ​​model selection and management service, and AWS SageMaker AI, the model building and training tool.

Interestingly enough, AWS is far from the only giant cloud provider to install Nvidia AI Factories. In October, Microsoft showed off its first of many future AI factories being rolled out across its global data centers to run OpenAI workloads. Microsoft did not announce at the time that these extreme machines would be available for private clouds. Instead, Microsoft highlighted how it was leaning on a host of Nvidia AI Factory data center technology to build and connect its new “AI Superfactories.” state-of-the-art data centers are being built in Wisconsin and Georgia.

Last month, Microsoft also outlined its data centers and cloud services that would be built in local countries to address the problem of data sovereignty. To be fair, the options also include ‘Azure Local’, Microsoft’s own managed hardware that can be installed at customer locations.

Still, it’s a bit ironic that AI is causing the largest cloud providers to invest so heavily in private data centers and hybrid clouds as if it were 2009 again.