On An Industrial Scale

The wonders that AI factories can power

With about 4 in 5 businesses employing AI in some form – some more than others – operations get complex. There’s a ton of data to store and it requires a great deal of security. Even for large corporations and governments, it all gets a bit overwhelming. Organizations like that need help managing all their AI capabilities, but they also don’t want to store their sensitive data too far away. That’s where AI factories come in. They provide dedicated infrastructure with state-of-the-art compute right in those organizations’ own existing data centers. It’s key to unlocking AI deployment at scale.

In the Zone

Achieving data sovereignty and meeting regulatory requirements for such organizations is the goal behind AWS’ AI factories. AWS is putting powerful tools to work in hitting these deliverables with speed, including Trainium3 chips – built for agentic, reasoning, and video generation applications – SageMaker AI and Bedrock for building and managing AI models. Bedrock helps select the right AI model for an individual use case, critical to the success of an endeavor but something that without expertise can involve a lot of time-consuming trial and error. Combine that with the computing power, dedicated data centers, and procurement, and organizations could be looking at years’ worth of work to scale up their AI operations. AI factories deliver all that utilizing existing data center space and power capacity. It’s like having a private AWS region, Amazon says.

“AWS AI Factories allow organizations to stand up powerful AI capabilities in a fraction of the time and focus entirely on innovation instead of integration,” Ian Buck, vice president and general manager of Hyperscale and HPC at Nvidia, said.

Yes, Nvidia, maker of some of the world’s fastest chips, is a collaborator in the AWS AI factories. The two companies are working together on an exceptionally ambitious project in Saudi Arabia they’re calling the first “AI zone.” Built in partnership full-stack AI ecosystem Humain, the AI zone will feature up to 150,000 chips, including GB300 GPUs.

“Through a shared commitment to global market expansion, we are creating an ecosystem that will shape the future of how AI ideas can be built, deployed, and scaled for the whole world,” Humain CEO Tareq Amin said.

Keeping Pace

Amazon is not the only fellow Magnificent 7 member that Nvidia is working with to establish AI factories. Microsoft is using Nvidia tech at its “AI superfactories” in Wisconsin and Georgia. Microsoft’s Azure global data centers sport more than 4,600 Nvidia GB300 computers using the ultra-high-speed InfiniBand network.

“Our collaboration with NVIDIA is built on driving innovation across the entire system and full stack, from silicon to services,” said Nidhi Chappell, corporate vice president of product management at Microsoft. “By coupling Microsoft Azure’s unmatched data center scale with NVIDIA’s accelerated computing, we are maximizing AI data center performance and efficiency, which is of paramount importance for our customers leading the new AI era.”

The sheer computing power these AI factories contain allow organizations to develop and manage digital twins and generative AI applications directly from the Azure cloud to factory floor or secure edge locations in real time. With more than 300 data centers in 34 countries and counting, Microsoft is out to remind industry and government leaders that it’s still a power player in cutting-edge solutions.

“Azure is proud to lead the way, supporting customers to advance frontier AI development,” Microsoft says.

Driving Physical AI

That kind of power is needed to bring physical AI to the forefront. For autonomous vehicles and smart cities to work, they need the kind of processing speed AI factories can deliver.

Hexagon Robotics’ AEON, a humanoid robot that uses its agility and spatial awareness to perform repetitive and hazardous tasks in industrial settings without stepping on the toes of human co-workers, relies on Nvidia’s full robotics stack on Azure. Wandelbots’ NOVA scalable smart factory software uses Nvidia robotics simulation framework to unlock critical workflows.

“Engineers and operators can model, test, and optimize complex operations faster – on-prem, in the cloud, or at the edge – with greater clarity and confidence,” said Prateek Kathpal, president of industrial at SymphonyAI, an enterprise leader in generative and predictive AI. “That’s the power of Vertical AI in practice.”

In Brownsville, Texas, a city-operated AI factory processes real-time data from cameras, sensors, and connected systems to enhance public safety, traffic control, and emergency response. The city launched a private 5G network last fall in partnership with NTT Data, transforming the border town from one of the nation’s worst-connected places to one of its best.

“Together, we’re building a digital foundation that positions Brownsville as a national model for next-generation community connectivity,” NTT’s Cathy LoDuca said.

It’s a testament to what AI factories can build.