How Qualcomm Dragonwing is putting industrial and physical AI to work

What you should know:

• IoT is evolving from connecting devices to enabling intelligent systems that can understand their surroundings and make real-time decisions.
• Qualcomm Dragonwing brings together compute, connectivity, AI, software and developer resources to help move industrial and physical AI from prototype to production through an open ecosystem.
• Vision systems and robotics work hand in hand to show how intelligence can operate close to where data is created while remaining connected to broader infrastructure and cloud capabilities.

IoT itself is being reimagined. Intelligence is becoming embedded in machines, robots, cameras, sensors and infrastructure, allowing these systems to understand their surroundings, respond in real time and work with broader infrastructure to interact with the world around them.

At the recent North America Dragonwing IoT Day, I shared why I believe this shift represents a significant opportunity for our industry and how Qualcomm is working with our ecosystem to help bring it to life.

How intelligence is changing IoT

Across factories, mines, stores, warehouses and other environments, we see a common pattern emerging: sites are becoming more connected, devices are becoming AI-enabled and workflows are becoming AI-aware. IoT is evolving from primarily connecting devices to enabling systems that can sense, understand data and support decisions in real time, often close to where that data is created. This is enabling additional automation capabilities and increasingly intelligent workflows where work actually happens.

This transformation is taking several forms. Human-facing intelligence is creating new ways for people to interact with devices. Machine-facing intelligence is giving systems greater situational awareness, combining signals such as vision and sensor data to understand what is happening around them and inform real-time decisions. And physical AI takes intelligence one step further, enabling robots, drones and autonomous machines to reason, sense, plan and act as they perform increasingly sophisticated tasks in the real world. 

Building the Dragonwing platform for this moment

Over the last two years, we've been on a journey to become a leading provider of compute and AI technology for the intelligent edge across key industrial and embedded verticals. That has required us to evolve not just our portfolio, but how we build products, develop software and engage with customers and developers. That work has included the launch of the Dragonwing brand, several acquisitions, a unified software architecture and Qualcomm Linux software stack, Arduino boards powered by Dragonwing processors and full-stack robotics.

We've also rebuilt our approach to IoT around this opportunity. Dragonwing brings together technology platforms, purpose-built products, software, developer enablement, solutions and services, supported by our foundational strengths in silicon and connectivity.

Just as importantly, we're working to make these technologies more open, accessible and easier to build with. The goal is to reduce friction throughout the developer journey without removing the deep customer engagement that has always been critical to the way we work.

How customers and developers can move industrial AI from prototype to production

Putting AI to work at scale requires more than powerful technology. Developers and customers need a path that makes it easier to access the right compute, software and tools, start building quickly, choose a platform and ultimately commercialize and manage products over long deployment lifecycles. Ease of access and ease of use are becoming foundational to how we scale Dragonwing IoT solutions.

That's why we're focused on bringing together developer-ready hardware, software, documentation, tools and ecosystem support. With Dragonwing IQ8 and Dragonwing IQ9 processors, for example, our approach includes EVKs, a common Linux foundation, centralized documentation and engagement through forums, Discord and GitHub, alongside our continued direct engagement with customers.

Arduino expands that opportunity even further, helping support development from accessible prototyping toward commercial deployment.

How vision AI helps machines see, understand and act

One area I'm particularly excited about is vision.

Cameras are the eyes of industrial and physical AI. As intelligence becomes embedded into buildings, factories, vehicles and robots, cameras are evolving from devices that simply capture or record what is happening into systems that provide situational awareness. They can help machines understand their surroundings, reason over what they see and determine what action to take in real time.

The Qualcomm Insight Platform extends this opportunity by connecting the camera chipset foundation through edge compute to cloud-based capabilities, supporting use cases across search, safety, security, operational analytics and more. 

What it takes to bring physical AI into the real world

Robotics takes this transformation another step forward. As robots evolve from inspection and mobility toward more complex interaction, manipulation and multi-agent fleets, the technology required to power them becomes increasingly sophisticated.

A robot isn't one computer. It effectively brings together different types of compute for sensing, action and reasoning, each operating at different speeds and with different requirements. Scaling physical AI requires these capabilities to work as a coordinated system that can sense, reason, plan, act and learn.

That's where we see an important role for Dragonwing: bringing together silicon, heterogeneous compute, sensing, connectivity, AI, control, data infrastructure and robotics software into an increasingly complete platform. 

Putting AI to work, together

The opportunity ahead is to transform operational technology (OT) with AI, supporting potential improvements in productivity, autonomy, efficiency and safety across industries. Our expanded portfolio spans foundational platforms and chipset products through full-stack solutions, helping give customers the flexibility to engage at the level that makes sense for their business, whether they want to build their own hardware or solve a specific operational problem.

But this isn't something any one company will accomplish alone. The pace of innovation is only accelerating, and bringing industrial and physical AI into the real world requires an ecosystem working together across hardware, software, tools and services to take innovation from prototype to production. That's what Dragonwing was built to help enable. 

Go Deeper

What does industrial AI look like in practice?

Industrial AI is taking shape across environments where machines need to perceive, understand and act on real-world conditions, from intelligent cameras and autonomous robots to connected infrastructure.

How is Qualcomm approaching physical AI and robotics?

For Qualcomm, robotics and physical AI are systems problems. Our end-to-end approach goes beyond silicon, bringing together hierarchical compute, advanced memory architecture, a robotics SDK that includes MoveIt, AI Ops, simulation, the data flywheel and applications. Together, these capabilities are designed to provide customers with an integrated solution or a platform on which to build and scale their own.