Over 130 Companies Adopt Intel® Series 3 Processor Family for Edge Devices

Intel announced today significant progress in Intel Robotics and Edge AI based on the Intel® Series 3 processor family, with over 130 edge AI and edge computing design projects currently underway. A notable example is the adoption by SensoryAI, which is deploying 'Ella,' the first multi-agent physical AI store for commercial service, and is transitioning to Intel architecture. Additionally, Intel announced 'OpenVINO™ Physical AI,' the industry's first open-source framework optimized for Intel CPUs, designed to solve deployment and scaling challenges at a low cost, enabling the robotics industry to move designs from development to production environments.

Dan Rodriguez, Intel Corporation Vice President and General Manager of the Edge Computing Division, stated, 'Physical AI models are revolutionizing robotics, but their implementation has been delayed by fragmented software stacks and the complexity of integration processes for each robot. By combining Intel® Core™ Ultra Series 3 processors with OpenVINO™ Physical AI, Intel provides an integrated, open, and scalable path that supports everything from the AI experimentation phase to the delivery of production-grade robots running high-performance inference via hardware acceleration.'

SensoryAI: Adopting Intel Architecture for the 'Ella' Multi-Agent Physical AI Store

As a demonstration of the effectiveness of Intel® Core™ Ultra Series 3 processors and Intel robotics solutions, SensoryAI has transitioned from a fragmented architecture using traditional CPUs and discrete accelerators in 'Ella' to the Intel® Core™ Ultra Series 3 processor platform, which can handle both real-time control and AI processing on a single chip.

As a result, they have realized a multi-agent physical AI store where three specialized AI agents—'Avatar,' 'Guardian,' and 'Ella Agent'—run simultaneously on a single SoC (System on Chip). This handles customer interaction, system operation, and store-level business intelligence in parallel, while a deterministic orchestrator controls the robots. This eliminates unnecessary components, reduces software complexity, improves ROI, and establishes a simple roadmap for future robot designs on Intel platforms. The latest demonstration of Ella is being showcased at the ongoing 'Computex.'

OpenVINO™ Physical AI: Bridging the Gap from Lab to Factory Floor

To help developers and operations teams transition smoothly from experimental models to large-scale deployment in real environments, Intel announced 'Physical AI Studio' and 'OpenVINO™ Physical AI' as part of its Robotics AI Suite.

With Physical AI Studio, developers can consistently perform data collection, model fine-tuning, optimization, and quantization, and seamlessly export pre-validated fine-tuned VLA (Vision-Language-Action) models for deployment in real environments. OpenVINO™ Physical AI is the industry's first open-source robotics library with a silicon-optimized inference runtime. Developers can transition robot control policies and multimodal models into practical robot systems in a consistent process while maximizing inference performance. Furthermore, it supports integration with open-source robotics model development environments like Physical AI Studio and LeRobot to achieve seamless model export.

The Missing Link in Robotics and Edge AI

For business operators facing labor shortages, rising costs, and intensifying competition, investment in robotics and automation to maintain production lines and improve productivity is an urgent priority. At the same time, robotics platforms are evolving from deterministic systems to autonomous physical AI systems that perceive, judge, and operate in the real world with millisecond-level precision.

Until now, large-scale deployment of physical AI models required highly customized pipelines to handle sensors, codecs, inference loops, and actuation for each robot. This forced the selection of over-specced dual-computing configurations that were costly to deploy and maintain.

By combining Intel® Core™ Ultra Series 3 processors and OpenVINO™ Physical AI, Intel provides an integrated hardware-software stack that allows customers to reduce Total Cost of Ownership (TCO), maximize code reuse across different types of robots, and efficiently scale and deploy robot fleets in a wide range of environments, including factories, warehouses, and retail stores.

Intel's robotics portfolio includes the latest software tools in the Intel Robotics AI Suite, robotics development kits provided by Intel partners, and Intel® Core™ Ultra Series 3 and Intel® Core™ Series 3 processors specifically designed for robotics and edge AI applications.

Availability:

OpenVINO Physical AI is available in preview on GitHub, with general availability planned for late 2026. Physical AI Studio is now available.