Hannover Messe remains a bellwether for industrial and manufacturing technology innovation with standout solutions displayed across Artificial Intelligence (AI), digital twins, industrial automation, and digital engineering. Within AI, the focus was Agentic AI—rather than copilots—with the prevailing sentiment to view agentic applications as coworkers who are tasked with a specific activity, rather than the chat-style copilots that permeated 2025. Physical AI was the other dominant theme, tugging on the thread of what’s possible versus available today with various demonstrations of Vision-Language Models (VLMs) and Vision-Language-Action (VLA) models used to enable autonomous automation. The biggest difference this year compared to past years is that many of the innovations the industry has been talking about—digital twins, control virtualization, and digital engineering—are ready for implementation rather than just talked about.
IMPACT
Small Change That Scales
|
Overall, the event drew 110,000 attendees and 3,000 exhibitors, which is slightly down from last year’s 125,000 attendees, most likely due to worker strikes making travel difficult. Siemens again had the largest booth, drawing tens of thousands of visitors each day. Metaverse reappeared, but the focus was more on the unification of data from different systems for a more complete, contextual, and comprehensive digital view of operations—less sci-fi immersive reality than in years past. Here, the industrial tech giant’s forthcoming Digital Twin Composer was a compelling demonstration of how pulling data together from different systems (Computer-Aided Design (CAD), Product Lifecycle Management (PLM), Application Lifecycle Management (ALM), simulation, Manufacturing Execution System (MES),…) can both ease and optimize change in production environments.
AWS showed a similar capability through the organization of robotic operations from different fleets to minimize the tolerance that comes from working in disparate systems—in this case, to save 9.5 seconds per robotic operation by better coordinating endpoints, which, at scale can mean thousands of additional minutes of uptime over the course of a year.
Zebra made its second appearance in Hannover with a dedicated booth and two of the best examples of an incremental innovation that yields a big change at scale:
- A label printer that prints directly onto an adhesive tape that can be placed directly on boxes/items without the need for backing paper—a small change that saves seconds per operation, minimizes waste, and allows more prints per roll because the paper is thinner without a release liner.
- An AI-enabled barcode scanner that batch-reads barcodes from a picture instead of individually scanning each barcode—a simple change that speeds barcode reading activities from 2 minutes to 20 seconds.
Tulip’s AI-tagged video analytics solution was another head turner. The company is best known for its composable MES solution for frontline workers. Now, it is building on its heritage by leveraging VLMs to create applications for customers based on video and audio media. Examples include creating assembly or service instructions based on a video recording and the ability to automatically tag videos so they are searchable, which can save hours of searching through production or security footage if trying to analyze an event.
RECOMMENDATIONS
"Where we're going, we don't need roads"
|
Industrial automation is about to have a full-circle renaissance moment. The main innovation focus 15 years ago centered on Machine-to-Machine (M2M) technologies—connecting and networking endpoints to perform analytics, predictive maintenance, and transform business models such as Rolls Royce’s now-famed power-by-the-hour scenario for jet engines. The last decade saw M2M evolve into Internet of Things (IoT), Industrial IoT (IIoT), and eventually intelligent automation powered and accelerated by AI. Interestingly, the current paradigm shift heralds back to M2M, except with an emphasis on machine-to-machine automation and control rather than simply communications, as evidenced by the rise of agents, multi-agent architectures, and Software-Defined Automation (SDA).
But AI alone doesn’t drive change. AI must be coupled with a strong data foundation (a comprehensive digital twin backbone), resilient networking infrastructure, and virtualized control architecture to truly drive new economies of efficiency. There will be a litany of unsuccessful agentic deployments in the next 12 months for suppliers that start with agents without working backward from the data foundation upon which they are built. A common pitfall will be in the use of VLMs and VLA models, which are shiny new objects in helping ease and accelerate automations but fall short if not linked back to design and engineering. Fortunately, there are a lot of great, low-cost, and easy-to-implement options to drive positive business impact before going all-in on AI-enabled automation.
