Intel announced early success in its robotics ecosystem at the opening of Computex in Taiwan early Monday (local time) while introducing the OpenVINO Physical AI framework to simplify deployment of next-gen robots.
The company said 130 companies are now testing or deploying applications with its Intel Core Ultra Series 3 processors, first introduced at CES in January. In one example, Sensory AI is moving to using Series 3 for Ella, a robot barista, making it the first multi-agent Physical AI store running in public commercial service, Intel said.
With its announcement of OpenVINO, Intel called out the difficulties of robotic deployments by developers. “Physical AI models are transforming robotics, but deployment has been slowed by fragmented software stacks and one-off integrations for every robot,” said Dan Rodriguez, corporate vice president for edge computing at Intel. “With Intel Core Ultra Series 3 and OpenVINO Physical AI, we provide a unified, open and scalable path from AI experimentation to production grade robots delivering hardware-accelerated, high-performance inference.”
Intel claimed OpenVINO Physical AI is the first open-source robotics library with a silicon-optimized inference runtime. This will add consistency for developers who take robot models from the experimental stage to working robotic systems while maximizing inference performance.
OpenVINO: Write code once
OpenVINO gives developers the ability to write code once to deploy on hardware that’s available and does an optimal job for the application, said Jack Gold, principal analyst at J Gold Associates. “OpenVINO is about optimizing deployment of AI compute across a variety of different systems. Using OpenVINO, devs can write code that is then deployed on CPU, GPU or NPU as required for the particular compute needs without the need to write code specifically for each compute engine,” Gold said.
“This capability enables a flexibility to write once, deploy on optimum hardware, depending on what’s available. While a lot of code is specific to the hardware, including CUDA for Nvidia GPUs, this approach attempts to make the code universally deployable based on the hardware you have implemented due to constraints of power, physical space and other needs. It also enables code reuse since different platforms of deployment have different processing systems designed-in.”
Core Series 3 and Ultra with Sensory AI and Ella
Gold said with Core Series 3 processors, Intel is targeting the x86 embedded systems marketplace for edge and AI devices. “Intel and x86 have a large share of this market and will continue to have this share since physical AI deployments are part of a greater need to do other compute functions in the physical/robotics world that are best handled by a CPU processing solution and not by dedicated GPUs,” Gold said in comments to Fierce.
“By downscaling the Core 3 to be more appropriate to embedded systems than are deployed in high power PCs, which is Core 3 Ultra, Intel provides a significant upgrade for those designs already using x86 designs and code bases. There is a large base of already-written x86 base code that can then be re-used.” (Note: Intel describes both its Intel Core Ultra Series 3 and Intel Core Series 3 processors as “designed specifically for robotics and edge AI applications.”)
Intel is undoubtedly positioning itself as a full-fledged Edge AI facilitator, with hardware and software, like never before, and calls its recently-introduced Robotics AI Suite a way for developers to onboard apps with lower R&D costs and the ability to scale up efficiently. The system is also open, which is how Intel hopes to take on entrenched CUDA.
At Sensory AI, the Ella robot barista retail system has been in place for six years with Crown Digital, in one example. On its website, Sensory AI notes that its infrastructure runs on Intel Core Ultra. “Intel is our silicon and edge-AI partner and Sensory AI is a title showcase at Computex 2026, presented as ‘Ella powered by Intel Core Ultra.’”
Intel said Sensory AI replaced a fragmented CPU plus discrete accelerator architecture in their Ella with a single Intel core Ultra Series 3 platform for both real-time control and AI. Intel wrote in a description: “The result is a multi-agent Physical AI store where three specialized AI agents—Avatar, Guardian and Ella Agent—run concurrently on a single SoC, handling customer conversation, systems operations and store-level business intelligence, while a deterministic orchestrator commands the robot.
“This eliminates an entire class of components, cuts software complexity, increases ROI, and creates a cleaner path to scale new robot designs on future Intel platforms. See what Ella’s brewing up at Computex in the Taipei Nangang Exhibition Center entrance hall.”