Why it matters now: For decades, the hardest frontier in industrial automation has been the handling of irregular, delicate, or unpredictable objects. Programmable logic controllers (PLCs) and robotic arms excel at speed and repeatability—but falter the moment a tomato is slightly overripe, a cable bundle shifts position, or a glass vial has a microscopic variation in wall thickness. That bottleneck has cost the global manufacturing sector billions in unrealized automation potential. At Automate 2026 in Chicago, a three-way alliance between PSYONIC, ABB Robotics, and NVIDIA demonstrated a pathway that could finally break the impasse.
Inside the PSYONIC–ABB–NVIDIA Collaboration
PSYONIC, a fast-rising bionics company, has developed a multi-articulated bionic hand that a human operator wears to perform real-world manipulation tasks. Every grasp, twist, and tactile adjustment the operator makes is captured as high-fidelity dexterity data—position, force, torque, and compliance—across multiple degrees of freedom. That data, once streamed into NVIDIA's AI infrastructure, becomes training fuel for industrial robots built on ABB Robotics' platforms.
ABB, one of the world's dominant PLC and industrial automation providers, brings the physical deployment layer. Its robot controllers and programmable logic systems are already embedded in factories worldwide. The question the trio posed at Automate 2026 was straightforward: what if those same PLC-driven arms could learn manipulation the way large language models learned language—through a massive, high-quality dataset?
The Collaboration at a Glance
| Partner |
Role |
Key Contribution |
| PSYONIC |
Data Originator |
Bionic hand captures human dexterity data across force, torque, and compliance vectors |
| NVIDIA |
AI Infrastructure |
Training pipelines, simulation environments, and edge inference for real-time robot learning |
| ABB Robotics |
Industrial Deployment |
Robot platforms and PLC control systems that execute learned behaviors in production environments |
The 'ImageNet Moment' for Industrial Robotics
The companies described their ambition as delivering robotics its “ImageNet moment.” In 2012, the ImageNet dataset catalyzed a revolution in computer vision—turning theoretical deep-learning models into practical, deployable systems virtually overnight. PSYONIC, ABB, and NVIDIA argue that human-generated dexterity data could do the same for physical manipulation. A bionic hand worn by a skilled technician generates training data that is orders of magnitude richer than anything a robot could collect autonomously, and far safer than trial-and-error learning on a live factory floor.
Analyst Insight: The convergence of human-in-the-loop data collection with industrial PLC ecosystems represents a structural shift. Rather than programming robots line-by-line for each new SKU, manufacturers could deploy general-purpose manipulation models that adapt in real time. For ABB, this strengthens the value proposition of its integrated control architecture—where PLCs, drives, and robots share a common data backbone.
Implications for PLC-Driven Smart Factories
This development lands at a pivotal moment for industrial control systems. The migration from rigid, pre-programmed automation to adaptive, AI-informed control is accelerating—and ABB's involvement signals that major PLC vendors see dexterity data as a competitive differentiator. In practice, an ABB robot controller receiving NVIDIA-processed manipulation models could adjust grip force, approach angle, and compliance in milliseconds, without a controls engineer rewriting ladder logic or structured text routines.
For factory operators, the payoff is tangible: reduced changeover times between product lines, lower scrap rates on fragile goods, and the ability to automate tasks—such as fresh food packing, pharmaceutical vial handling, and electronics assembly—that have stubbornly resisted robotic penetration.
FAQ: What This Means for Automation Engineers
Q: Does this replace traditional PLC programming?
No. PLCs remain the backbone of safety, sequencing, and deterministic control. AI-driven manipulation layers sit on top, informing motion profiles and adaptive behaviors that the PLC orchestrates.
Q: Is specialized hardware required?
ABB's existing robot controllers can interface with NVIDIA's edge AI modules. PSYONIC's bionic hand is a data-collection tool—not required on the factory floor for inference.
Q: When will this reach commercial deployment?
The companies have not announced a firm timeline, but demonstrations at Automate 2026 suggest pilot programs with early-adopter manufacturers are underway.
Automate 2026: A Bellwether for Adaptive Automation
Automate 2026 in Chicago drew record attendance, with dexterity and adaptive manipulation emerging as the show's dominant themes. The PSYONIC–ABB–NVIDIA demonstration stood out not for flashy hardware but for its data-centric thesis: that the path to truly flexible robots runs through human expertise, captured at scale and translated by AI into executable control signals on industrial PLC platforms.
Market Trend: The global industrial robotics market is projected to grow at a compound annual rate exceeding 10% through 2030, with the fastest growth concentrated in collaborative and AI-capable systems. Dexterity—not speed or payload—is now the primary barrier to adoption in sectors from agriculture to medical devices. Alliances like this one suggest the barrier is beginning to crack.
The Road Ahead
PSYONIC, ABB Robotics, and NVIDIA have not disclosed the scale of their data-collection efforts, but the logic is compelling: every hour a human operator wears a bionic hand, the dataset grows richer. Every inference cycle on an NVIDIA GPU turns that data into a more capable manipulation policy. And every ABB robot running that policy brings adaptive dexterity closer to being a standard feature of industrial PLC systems—not a research curiosity.
For the industrial automation community, the message from Chicago is unmistakable. The robots are learning. And this time, they are learning from us.
