Intel has taken a strategic step by venturing into the realm of physical artificial intelligence, a move that marks its return to the robotics market, a sector it had abandoned years ago due to financial difficulties. This decision is not isolated but part of a more ambitious plan to position AI at the edge, where devices can run AI models locally without relying exclusively on the cloud. In a context where many devices lack local computing capabilities, Intel aims to fill that gap with efficient and powerful processors.

Intel Core Ultra Series 3 Processors: The Heart of the New Robotics

The Intel Core Ultra Series 3 processors, originally designed for laptops, have been adapted for robotics and edge applications. Intel has achieved a level of energy efficiency that allows for long battery life, making them ideal for portable and handheld devices. According to the company, these chips are already present in 130 AI edge and robotics designs, demonstrating significant early adoption.

A notable case is SensoryAI, which uses Intel technology to power Ella, a robotic barista created by Crown Digital. This robot not only makes coffee but integrates multiple AI agents on a single chip. The main agent "Avatar" serves customers, while "Ella" reasons and executes tasks. If errors arise, such as misinterpreting an order or managing stuck cups, a "Guardian" agent intervenes for recovery. All of this runs on a single Core Ultra Series 3 silicon piece.

Integration of Functions in a Single Chip: The Leap Toward Efficiency

One of Intel's key innovations is the ability to integrate multiple robotic functions—such as computer vision, real-time controls, graphics, and motion—into a single chip. Previously, these functions were distributed among different cores within a chip, increasing complexity and energy consumption. With the new processors, Intel simplifies the architecture, improving performance and reducing latency.

This advancement is possible thanks to Intel's latest manufacturing technologies, which allow for the production of advanced chips for robotics with a high degree of integration. The company has demonstrated these robots at the Computex fair in Taiwan, and even shared a video of a humanoid robot on its social media, underscoring its commitment to physical AI.

Implications for Businesses and IT Professionals

Intel's focus on physical AI and edge computing has profound implications for businesses. By decentralizing AI processing, organizations can reduce reliance on the cloud, decrease latency, and improve data privacy. This is especially relevant in sectors such as manufacturing, logistics, and customer service, where robots and autonomous devices require real-time responses.

Moreover, the integration of multiple AI agents on a single chip opens the door to more autonomous and resilient systems. As mentioned in our article on agentic AI, the ability to coordinate multiple agents within the same hardware is crucial for building robust operations platforms.

For IT professionals, this means they will need to become familiar with edge computing architectures and optimizing AI models for specific hardware. The trend toward physical AI also poses challenges in terms of security and maintenance, as discussed in the problem of accountability in AI agents.

A Growing Ecosystem

Intel is not alone in this race. Other companies like NVIDIA and AMD are also developing chips for edge AI, but Intel's bet on robotics and single-chip integration gives it a competitive advantage. The company has indicated that its long-term strategy includes collaborations with robot manufacturers and software developers to create a complete ecosystem.

At ForgeNEX, we have analyzed how the implementation of generative AI in workflows (see article) can benefit from this type of hardware, allowing complex models to run locally without sacrificing performance.

Intel's return to robotics is not only a sign of its financial recovery but also an indicator of where the industry is headed: toward an AI that not only thinks but also acts physically in the real world.

Original source: ComputerWorld. Analysis and adaptation by ForgeNEX.