From Co-Pilots to Humanoids: Mapping the Entire Industrial AI Battlefield

For several years, I’ve been writing about the "Industrial AI Wars"—a high-stakes battle waged on multiple fronts to reshape industry with intelligent software and hardware, which I updated again in “The Industrial AI Revolution: Navigating the AI Wars in 2025.”

This is a war being fought on multiple, concurrent fronts.

Lately, one of those fronts has captured the media’s and the market’s imagination: the "Industrial Robot Wars."

This is the dynamic, high-stakes battle for Physical Intelligence, where a new generation of robots is learning to perceive, reason, and act in the physical world. This "war for motion" is a paradigm shift, and to help our clients navigate it, my colleague Patrick Arnold is leading our definitive research. He is applying our foundational Physical Intelligence Taxonomy to map the "body"—the physical platforms, hardware stacks, and operational software that will form the new industrial workforce. His work is essential, as you can’t have a revolution without the new "vessels" of automation.

But this "war for motion," critical as it is, is only one front in the larger Industrial AI Wars.

What about the "war for engineering," where AI co-pilots are helping design the next generation of products and factories? What about the "war for operations," where autonomous AI models are optimizing a chemical process in real time or predicting failures in a fleet of assets? These are the battles being fought with "bits" and "atoms," and they are just as consequential as the one being fought with "motion."

A Market Clouded by AI Hype

The problem is, the vendor landscape is making it impossible to see the full battlefield clearly.

The "AI" gold rush is on, and every vendor—from enterprise software to automation incumbents to cloud hyperscalers—is now an "AI vendor." They are all in a frantic push for their slice of a very large pie, and in the process, they are "clouding" the landscape with marketing hype.

This lack of discipline and clarity is a critical problem for industrial clients. How can you develop a strategy when you can’t tell what’s real?

For example:

• Is the "AI" in your process optimization tool just a re-badged statistical model, or is it a true Physics-Informed Neural Network (PINN) that actually understands the underlying chemical process?

• Is that "Autonomous Operations" platform simply a better dashboard for your operators, or is it a true Foundation Process Model trained on real-time OT data, capable of running the plant in a closed-loop, autonomous mode?

• Is that new "Industrial Co-pilot" just a thin wrapper on a general-purpose LLM, or is it a purpose-built model trained on the complex logic and physics of engineering data, CAD schematics, and PLC code?

• Is the "AI brain" for that new humanoid robot just a collection of preprogrammed, fixed-function skills, or is it a foundational Embodied Intelligence Model—perhaps an early "Large Behavior Model" (LBM)—that can actually learn and reason about new physical tasks in unstructured environments?

A Map for a New Battlefield

Right now, vendors are using the same "Industrial AI" umbrella to cover all of these—and more. That is not sustainable for buyers or a healthy market.

This is why my research is focused on providing the master map for the "brain" in all its forms. We need a new framework that brings discipline and clarity to this chaotic market, allowing industrial leaders to understand, categorize, and plan their AI investments.

That framework is the ARC Industrial AI Models Taxonomy.

The "Central Nervous System" for All Fronts

It’s the map that shows how all these fronts are related. And, of course, all these battlefronts— from Patrick's physical robots to my engineering co-pilots—have one nonnegotiable prerequisite: the "central nervous system" that is the Industrial Data Fabric.

This blog series will lay out this new map. We’ll show how it brings clarity to the entire Industrial AI Wars, not just the robotics front. We’ll define each of the four levels of intelligence and show how Physical Intelligence (Level 4) is the capstone of a full-stack AI strategy that begins with knowledge, "bits," and "atoms."

In my next post, I’ll dive deeper into why the Industrial Data Fabric is the nonnegotiable "cost of entry" for competing in—and winning—any of these AI wars.

Engage with ARC Advisory Group

For ARC Advisory Group recommendations for Navigating the AI Wars—including the Industrial Robot 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 my fellow Analysts and I, at ARC Advisory Group to find out more about our Executive Insights Service for Industrial organizations, and Industrial AI Insights Service for Vendors.