Representative End User Clients
Representative Automation Clients
Representative Software Clients
In our first ARC Advisory Group SPARC podcast in this series with AWS, I had the pleasure of speaking with Steve Blackwell, Head of the Manufacturing Center of Excellence at Amazon Web Services (AWS), about the transformative shift towards Software Defined Manufacturing (SDM) in the Era of Agentic AI. We explored how Agentic AI is becoming the "brain and central nervous system" of the future factory, orchestrating complex digital workflows across the enterprise.
In this second episode, we bring that vision down to the factory floor. We're moving from digital intent to physical action—from the brain to the "muscles" of the operation. Steve and I discuss the physical embodiment of SDM through robotics, exploring AWS's evolving strategy and the rise of a new, more capable class of automation: Physical and Embodied AI.
Listen or Watch
You can listen to or watch our full conversation here:
Listen on the Digital Transformation Podcast on BuzzSprout:
Watch on ARC's YouTube Channel:
For those who prefer to read, I’ve distilled our conversation into the key insights and recommendations below.
Key Insights and Recommendations
1. AWS’s Strategic Pivot: From Monolithic Service to Composable Architecture
A key development in the industrial robotics space is AWS's decision to sunset its developer-focused AWS RoboMaker service. Rather than a retreat from robotics, this signals a strategic pivot toward a more flexible and powerful composable architecture. AWS is empowering manufacturers to build custom, resilient, and deeply integrated automation solutions by combining foundational services like AWS IoT for connectivity, AWS Batch for simulation, and AWS Step Functions for workflow orchestration. This approach acknowledges that every factory is different and that a one-size-fits-all product is less effective than a robust toolkit for building tailored solutions.
Steve Blackwell (AWS): "As this world evolves, we're always looking at how to support our customers and partners better. The technology and the needs of our customers are changing, and that's why we've pivoted from where we saw the industry four or five years ago to where it is now."
Colin Masson (ARC Advisory Group): "This pivot to a more composable approach better addresses the core enterprise challenges of integrating and managing heterogeneous, multi-vendor robot fleets, rather than being prescriptive about how to do it."
2. The Next Frontier: Defining Physical and Embodied AI
The conversation quickly moved beyond traditional automation to define the next frontier: Physical and Embodied AI. Unlike rigidly programmed robots that excel at repetitive sequences, these new systems are designed to perceive, reason, and act within dynamic, real-world environments. Steve defined Physical AI as an autonomous system that understands the world around it, while Embodied AI is the deep integration of that intelligence into the physical device itself, allowing it to perform autonomous tasks.
Steve Blackwell (AWS): "I describe Physical AI as an autonomous system, like a robot or an AMR, that understands the real world around it and uses AI to perform tasks. Within Physical AI, there's a subset called Embodied AI, which is the integration of AI into those physical devices, be it AMRs, robots, and so forth."
Colin Masson (ARC Advisory Group): "Traditional robots are pre-programmed and can be very efficient, but we're really talking about a move to systems that can perceive, reason, and act in a more dynamic way."
Sidebar: Amazon's Robin—Embodied AI in Action
A prime example of Embodied AI discussed in the podcast is Amazon's "Robin" robotic arm, one of the most complex stationary robot systems Amazon has ever built. Deployed in fulfillment centers, Robin uses computer vision, AI, and machine learning to identify, pick, scan, and sort packages from a jumbled conveyor onto a drive robot.
What makes Robin special is its ability to operate in a constantly changing environment. It doesn't just follow a preset script; it perceives the scene, understands different package sizes and shapes (even when they are overlapping), and decides which one to grab in real-time. This capability, which has allowed Robin to handle over one billion packages, showcases the power of training AI models to give robots the intelligence to perform complex, autonomous tasks that were previously impossible for industrial automation.
Learn more about how Robin works at Amazon Science:
Amazon’s robot arms break ground in safety and technology https://www.amazon.science/latest-news/amazon-robotics-see-robin-robot-arms-in-action
3. Industrial DevOps: Bringing Modern Software Practices to the Factory Floor
A core promise of Software Defined Manufacturing is applying modern software development practices to the operational technology (OT) world. We discussed the concept of "Industrial DevOps," where practices like cloud-based version control, automated testing, and CI/CD pipelines are used to manage PLC and robot code. This represents a tangible, high-value first step for manufacturers, solving real-world problems like disaster recovery, ensuring fleet-wide consistency, and reducing downtime associated with code fixes. This IT/OT convergence is critical for building the agile, resilient, and secure factories of the future.
Steve Blackwell (AWS): "What we're seeing now is that AI agents can be used to orchestrate multiple steps across tasks. For example, integrating the machine learning model that has identified a defect with the MES system where that non-conformance needs to be registered. This is where we're starting to see AI agents really come in."
Colin Masson (ARC Advisory Group): "A core promise of Software Defined Manufacturing is bringing modern software concepts and practices to the factory floor, an area that has perhaps fallen a little behind."
4. The Power of the Ecosystem: No One Builds the Future Alone
Realizing the vision of an autonomous factory is too complex for any single company. It requires a robust ecosystem of partners bringing their unique expertise to the table. We discussed how AWS is focusing on enabling this ecosystem, working with industrial automation leaders like Siemens and Rockwell, AI powerhouses like Nvidia, and the full spectrum of robotics OEMs. By providing the foundational cloud services and promoting interoperability through standards, AWS allows manufacturers and their partners to assemble the "Lego pieces" needed to build these sophisticated Physical AI systems securely and at scale.
Steve Blackwell (AWS): "We're very conscious that a manufacturer is not likely to have the skill sets to achieve this from end to end by themselves; they're going to need their partners. Whether it's an industrial automation partner like Siemens or Rockwell, a robotics partner like FANUC, KUKA, or ABB, or the AI partner ecosystem, they all need to come together."
Colin Masson (ARC Advisory Group): "It's important to understand how the ecosystem is working together to ensure this happens responsibly, safely, and productively as we move forward, because that's always the fear."
This has been a fascinating look at how software is truly meeting the physical world. In our next episode, I'm excited to dive into the "brain" of the operation with Steve Blackwell of Amazon Web Services (AWS): the digital twin. We'll explore how these high-fidelity virtual models, powered by simulation and a rich digital thread, are becoming the essential engine for achieving true, closed-loop autonomy in manufacturing.
Diving Deeper
As we venture further into topics like building robust data infrastructures and modernizing architectures to effectively infuse AI, I highly recommend readers explore some of my existing research:
My blog series on "Assembling Industrial-grade Data Fabrics."
My series on "The Rise of A2A: Completing the Industrial AI Protocol Stack with OPC UA and MCP."
The article, "Core Capabilities of the Industrial-grade Data Fabric: Powering AI Infusion and Modernization," which is the sixth post in the Data Fabric series and delves into the essential solution services these fabrics enable.
The new series on Physical Intelligence from ARC Advisory Group Analysts Patrick Arnold, Colin Masson and Craig Resnick, "Bridging the Digital and Physical: A New Era for Industrial Robotics"
These ARC Advisory Group articles offer additional context on AWS and NVIDIA in the industrial sector:
AWS Strands: Weaving a More Collaborative Future for AI Agents in the Industrial Stack (https://www.arcweb.com/blog/aws-strands-weaving-more-collaborative-future-ai-agents-industrial-stack)
The "Brain" for the Autonomous Factory: Why NVIDIA Jetson Thor Marks a New Era for Industrial AI and Robotics (https://www.arcweb.com/blog/brain-autonomous-factory-why-nvidia-jetson-thor-marks-new-era-industrial-ai-robotics)
NVIDIA Jetson Thor Ushers in the Age of Agentic AI-Powered Robotics (https://www.arcweb.com/blog/nvidia-jetson-thor-ushers-age-agentic-ai-powered-robotics)
Stay Connected
The dialogue on industrial AI and digital transformation is constantly evolving. To stay informed and hear more insights from industry leaders, we invite you to subscribe to the ARC Advisory Group's Digital Transformation podcast series.
We believe the best conversations include diverse perspectives. If you are an innovator in this space and would like to contribute to a future discussion, please reach out to Colin Masson at ARC Advisory Group.
Engage with ARC Advisory Group
For ARC Advisory Group recommendations for Navigating the AI Wars, Closing the Digital Divide by Embracing Industrial AI, assembling your Industrial-Grade Data Fabric, and governing and guiding major decisions about enterprise, cloud, industrial edge, and AI software, please contact Colin Masson at [email protected] or set up a meeting with me, or my fellow Analysts at ARC Advisory Group to find out more about our Executive Insights Service for Industrial organizations, and Industrial AI Insights Service for Vendors.