New AI chips are helping robots work faster and without bulky hardware, powering coffee-making machines, humanoid assistants, and factory automation systems.
Intel is pushing deeper into physical AI with a new generation of robotics systems powered entirely by Intel Core Ultra Series 3 processors, eliminating the need for discrete GPUs in deployed machines.
The system runs three AI agents simultaneously on a single Intel chip, handling customer interaction, operational reasoning, and machine monitoring without relying on cloud infrastructure.
The launch highlights Intel’s broader strategy to position its Core Ultra Series 3 platform as a scalable edge AI solution for robotics developers looking to reduce system cost, power consumption, and thermal complexity.
Ella represents one of the clearest examples of that transition. The robot, designed for hospitals, cafés, and retail environments, uses Intel’s integrated CPU, GPU, and neural processing unit (NPU) architecture to run vision, language, and motion-control workloads locally. Tasks that previously required large, power-hungry graphics cards can now run on a single system-on-chip platform.
Since adopting Intel Core Ultra Series 3, Sensory AI has transitioned Ella entirely to an Intel-only architecture. According to the company, the new platform enables agentic AI deployment at the edge while lowering total cost of ownership and simplifying maintenance.
The robot’s AI stack is divided across three specialized agents. The Avatar Agent manages customer communication, the Ella Agent analyzes operational trends and store-level patterns, and the Guardian Agent oversees diagnostics and recovery processes. If the system encounters a problem — such as cups sticking together — the agents coordinate detection, customer notification, and corrective robotic action autonomously.
“We used to have an architecture of an Intel CPU with a discrete GPU doing some of the workloads, but it was expensive,” says Tan, now founder and CEO of Sensory AI. “The GPU costs more than the entire system. I realized I just can’t spend that much. I have to build a system with the ROI of a cafe. Developers need to look at the total cost of ownership and the real-world deployment when you’ve already trained your models.
Click here for the original announcement.
