Intel looks to accelerate AI at the edge with new Edge Platform

Pallavi Mahajan, corporate vice president and general manager for the network and edge group at Intel

BARCELONA — Intel is attempting to make it easier for customers to build, run, secure and manage AI applications at the network’s edge with the launch of the new Intel Edge Platform, announced at this year’s Mobile World Congress here.

The company says the Platform provides several ways for companies to ease the process of building, running and maintaining AI-centric applications at the network’s edge. And it’s part of a broader attempt by the chipmaker to change the understanding of the role of AI at the edge.

Pallavi Mahajan, corporate vice president and general manager for the network and edge group at Intel, told here that when customers think AI, they think of the heavy lifting required for generative AI — the need to “throw in a GPU cluster to make it happen.” While it might be underestimated in the current GenAI mania, the good news is that most edge AI applications use it for inferencing. That will run handily on existing edge hardware with the help of an inference runtime like Intel’s own OpenVINO, which is part of the platform.

“It’s a way to go out on your existing infrastructure and run more AI inferencing. You don’t want to have to run three or four different solutions separately. You want to consolidate that, make it simpler to manage, and get your best total cost of operation,” Mahajan said.

Intel is looking at tasks like self-checkout, automated stock monitoring, personalized in-store advertising in retail, defect detection and predictive manufacturing maintenance, and various smart city applications as likely use cases. But in typically Intel fashion, it’s not going it alone — the platform will be built out and built upon an ecosystem of partners ranging from SIs to OEMs to ISVs. The goal, Mahajan said, is to avoid a “walled garden” approach that makes it easier to deploy, operate, and manage applications at the edge but does so in a way that limits what can be run on the platform and who can build on top of it.

“We don’t want our customers to feel like they’ve locked into one vendor, and they don’t have the choice to go out and experiment with different solutions,” she said. “Our philosophy here at Intel has been open first. So, how do we also create an ecosystem of partners? We feel that we don’t have to solve all the problems. We have to use all the all the brilliance in the world around us. So, [it’s about] nurturing and harvesting this ecosystem of partners.”

While that ecosystem approach may focus on ISVs on the developer side, it may prove interesting to the growing number of MSPs who are “productizing” their offerings in the AI era and are looking more and more like ISVs. Mahajan said while many of Intel’s example use cases envision, for example, retail chains with hundreds or even thousands of locations to bring online with new applications, the same fundamentals apply to the smaller retailers who may be having their applications built and managed by an MSP. 

“What we’re providing is an easy toolkit, where you can deploy in no-code or low-code, or where you can bring in your own code, and you don’t have to be a data scientist to actually build an application,” Mahajan said.

The platform consists of four layers. At the base, as with most things Intel, is the silicon on which applications run. Immediately above that, there’s an infrastructure layer that builds in security and manageability, features that make it easier and more efficient to run and maintain the applications running on the platform. The AI itself lives immediately above that in an application layer, designed to get the best power and performance out of the underlying hardware, and on top of that is what Intel calls the industry solutions layer. 

This final layer will be primarily the playground of the company’s edge ecosystem partners, which currently include hardware vendors such as Lenovo, major GSIs, and ISVs. However, Mahajan stressed Intel’s interest in keeping it an open environment, suggesting opportunities for MSPs and other types of solution providers that are developing, deploying, or managing edge applications for customers.

Robert Dutt

Robert Dutt is the founder and head blogger at He has been covering the Canadian solution provider channel community for a variety of publications and Web sites since 1997.