Intel has unveiled an initiative to streamline AI adoption at the edge across industries such as retail, manufacturing, smart cities and media by simplifying integration with existing infrastructure.

“Our customers want to expand the use of AI in their existing infrastructure and workflows at the edge, ensuring they meet their total cost of ownership and achieve power and performance goals,” said Dan Rodriguez, Intel corporate vice president. “With decades of experience at the edge, we’re taking our edge AI offerings and support one step further with Intel AI edge systems, edge AI suites and open edge platform to accelerate the delivery of AI across the ecosystem.”

Edge AI is becoming an essential driver of enterprise innovation. Gartner forecasts that by the end of 2025, half of enterprise-managed data will be processed outside traditional data centres or the cloud, in places such as retail stores, manufacturing plants and healthcare facilities. By 2026, at least half of edge computing deployments will involve machine learning.

With more than 100,000 real-world edge implementations with partners, many leveraging AI today, Intel says it understands the challenges of edge AI. These vary significantly by industry, with each having distinct performance and power needs. What works for cloud providers is not suitable for enterprises at the edge, which need to maintain existing platforms and software, while integrating AI, to get the best total cost of ownership (TCO) and power efficiency.

Unlike large data centres with dedicated AI infrastructure, edge AI deployments must seamlessly integrate into pre-existing IT systems in space-constrained, low-power and cost-sensitive environments.

Intel’s AI edge systems, alongside edge AI suites and the open edge platform, address these challenges by building on Intel’s technology foundation at the edge. These efforts can empower the ecosystem to bring edge AI to market faster and more efficiently.

The open edge approach allows Intel to deliver better end-to-end performance and overall TCO consistently in various key industries. In an edge AI video analytics use case, tera operations per second (TOPs) alone don’t deliver real-world performance needs. When comparing Intel Core Ultra processors to leading AI competition, although the competition may lead in TOPs, Intel says it delivers up to a 2.3-times increase in end-to-end pipeline performance, and up to fivefold better performance per dollar.

While many edge deployments today incorporate AI using traditional machine learning and computer vision, Intel’s latest initiatives are designed to accelerate the deployment of AI applications. Through a robust ecosystem of partners, Intel enables enterprises to tackle industry-specific problems and drive innovation in edge AI deployments.

Intel’s edge systems can speed up AI deployment at the edge. Builders – original equipment manufacturers (OEMs) and original design manufacturers (ODMs) – have access to standardised blueprints, benchmarks and verification tools optimised for edge AI use cases. These resources help users and providers easily configure systems to meet the performance needs of use cases such as vision AI or generative AI. With a range of power levels, sizes and performance options, these can ensure optimal integration of hardware and software.

Edge AI suites are open, industry-specific AI software development kits (SDKs) for independent software vendors, system integrators and builders. These suites simplify the creation of custom AI offerings for various industries by providing curated reference applications, sample code and benchmarks to accelerate application development. Currently, Intel offers four suites, optimised for retail, manufacturing, smart cities, and media and entertainment.

The open edge platform is a modular, open-source platform that simplifies the development, deployment and management of edge and AI applications at scale. With cloud-like simplicity, this platform allows ISVs, builders and operating system vendors to integrate software components efficiently and leverage performance optimisations from Intel’s latest software. Partners can deploy containerised workloads on remote edge devices without the need for on-site visits and manage these deployments with tools such as Intel vPro and active management technology, enabling collaboration and innovation across the software ecosystem.

More details on Intel’s edge AI portfolio can be found at www.intel.com/content/www/us/en/software/edge-platform.html.