With chips designed to accelerate AI now available across the company’s portfolio of PCs and edge systems, Intel is focusing on building an ecosystem around AI at the edge.

The chipmaker has announced that its OEM and ODM partners will launch a series of edge systems designed for AI applications later this year. It has also introduced new tools to build and validate AI edge applications, as well as software to assist in building and deploying edge-based AI systems and applications at scale.

“We’re in the era of AI at the edge, and it has unique needs,” said Dan Rodriguez, corporate vice president and general manager of the edge computing group at Intel. “It’s all about delivering AI in the existing footprint with existing workflows.”

Edge-based applications pose a unique challenge for AI in that the typical cloud-based AI approach of throwing racks of servers at compute-heavy workloads doesn’t work when those systems have to sit on a manufacturing floor or a show floor, Rodriguez said. Typically, applications have to run well on the same footprint where existing compute and media applications are running.

“There are all kinds of legacy infrastructures in place and operational complexities,” Rodriguez said. “We’re working with our customers and the ecosystem to ensure all of these workloads, including AI, can function together with the right TCO and the right user experience.”

AI-accelerated edge systems are central to the strategy, with Intel engaging its vendor partners to develop edge systems that address various computer vision and generative AI applications desired by customers. Rodriguez stated that Intel is assisting its hardware partners in moving more quickly and confidently with their new systems, while ensuring a range of configurations for power usage, performance, and physical size to meet the diverse requirements of different edge installations. All systems will undergo benchmarking for performance (including horsepower and power consumption) and size, aiding customers in their selection process.

Rodriguez acknowledged that some AI use cases may benefit from a separate GPU but said that for “a healthy majority” of edge use cases, “a CPU with integrated AI acceleration is great.”

Systems benchmarked by the chipmaker will carry an Intel Edge AI Systems brand.

Rodriguez stated that the company will also introduce Edge AI Suites, which are collections of sample code, demos, benchmarking, and tools designed to assure customers that Intel-based systems will perform effectively for edge AI applications. The initial Edge AI Suites will focus on manufacturing, retail, smart cities, and media entertainment. Rodriguez added that these resources will assist developers in creating “the best applications and maximizing our technology.”

Rodriguez said the suites will be expanded over time within the areas of focus. For example, at launch, the retail Edge AI Suite will focus on self-checkout systems but will expand to loss prevention and order accuracy systems in the future. Rodriguez also said additional suites may be added in other verticals that Intel focuses on for its edge business, including healthcare and government.

At the same time, the company will debut the Open Edge Platform, an open-source initiative that assists with the deployment and management of AI-based edge applications on its systems. Rodriguez stated that the platform, which is set to launch early in the second quarter of this year, will include security, provisioning, patch management, and other management tools based on the company’s vPro management technology. The platform will enable partners to build and deploy workloads on edge devices remotely and will feature solution development libraries.

“All of this helps the ecosystem move faster,” Rodriguez said.