Cloudflare and Microsoft Partner to Optimize AI Model Location

Cloudflare, a connectivity cloud company, and Microsoft are now collaborating to enable businesses to run artificial intelligence (AI) models in the most suitable locations for their needs.

This partnership aims to seamlessly deploy AI models across devices, network edges and cloud environments, the companies said in a Wednesday (Sept. 27) press release. 

The collaboration leverages ONNX Runtime, a standardized solution that allows the deployment of AI models regardless of the environment, according to the release. 

This means that AI models can be run wherever it makes the most sense, whether on devices, edge networks or in Azure’s centralized cloud platform, the release said. By utilizing this three-tier architecture, businesses can maximize the benefits of both centralized and distributed computing models. 

AI model training requires significant computational and storage resources, making centralized cloud platforms the best environment for intensive calculations, according to the press release. 

However, inference tasks can be performed in more distributed locations, the release said. For example, tasks like autonomous vehicle braking at the sight of a pedestrian can run on the physical device for the lowest possible latency.

Edge networks, which are in close geographical proximity to end users and devices, provide an optimal balance of computational resources, speed and data privacy, per the release. For example, a security camera system could utilize edge networks for object detection, overcoming the resource constraints of the device without the latency of sending data to a central server for processing.

The collaboration between Cloudflare and Microsoft aims to build the infrastructure for AI traffic and tasks, tailoring AI inference to the specific needs and demands of each organization, according to the press release. Businesses will have the flexibility to choose the best location for their AI tasks, based on factors such as speed, accuracy, energy, connectivity bandwidth and regional localization requirements.

Cloudflare’s global network, combined with Microsoft’s experience in training and deploying advanced AI workloads through Azure, will empower businesses to boost efficiency across the entire AI lifecycle, the release said.

PYMNTS Intelligence has found that generative AI has taken off exponentially, with new companies adopting the technology almost every day. These companies are aiming to use the technology in applications that will save time and labor in every field, from healthcare to finance, according to “Preparing for a Generative AI World,” a PYMNTS and AI-ID collaboration.

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.