A new industrial AI partnership aims to connect engineering blueprints with real-world machine behavior, closing a loop manufacturers have been trying to close for decades. 

by Zaara Abbas, Digital Media Manager at Tech Revolt

[For more news, click here]

There is a problem that most manufacturers know intimately but rarely discuss publicly. The factory that gets built is never quite the factory that was designed. Equipment degrades in ways simulation does not predict and maintenance schedules drift from engineering assumptions while production data sits in systems that were never intended to talk to each other. The result is a persistent gap between engineering intent and operational reality, one that costs manufacturers in downtime, margin, and agility every day. 

Siemens and IFS announced a strategic partnership on 29 June 2026 aimed directly at that problem. The two companies are combining Siemens’ digital twin technology and manufacturing execution capabilities with IFS’s enterprise asset management and field service platforms to create what they describe as a closed-loop industrial AI system, one that connects engineering design with real-world asset performance and feeds what it learns back into better design. 

The gap that costs manufacturers most 

The structural problem the partnership addresses is not new, but it has become more expensive to ignore. Manufacturers are under growing pressure to increase output from existing assets, manage margins, extend equipment life, and respond more quickly to changing conditions, all while operating with production, maintenance and supply chain systems that were never designed to share data with each other. 

The consequences are tangible. Unplanned downtime remains one of the largest controllable cost drivers in industrial operations. Maintenance schedules that cannot draw on real engineering data tend to be either too conservative, wasting resources, or not conservative enough, resulting in failures. Supply chain disruptions that a more adaptive system might absorb instead propagate through the entire production floor because the systems involved cannot communicate fast enough to respond. 

Industrial AI is increasingly positioned as the mechanism that bridges these gaps, but most AI systems in industrial environments are built on generic models that have not been trained on the specificity manufacturing demands. In environments where a small error rate is unacceptable because decisions affect safety, regulatory compliance, and costly physical assets, accuracy and reliability are not features. They are requirements. 

What each company brings 

The partnership is structured around complementary data domains. Siemens contributes its Digital Twin technology, which captures engineering, simulation, and manufacturing context: what a product or production system was designed to do, how it was modelled, and how it should perform. IFS contributes service history, asset behaviour records, and operational lifecycle data: what the product or system has done in the field, how it has degraded, how it has been maintained, and where the gaps between design intent and real performance have opened up. 

Together, the two companies plan to build what they are calling a closed-loop Digital Twin grounded in both sets of data simultaneously, auditable across design, simulation, service records, and factory execution, and designed to operate at industrial scale rather than in controlled pilot conditions. The system is intended to let manufacturers not just monitor the gap between design and reality, but use that gap as an input, feeding field performance data back into engineering and production planning so that future designs and maintenance strategies are informed by what actually happens, not only by what was originally assumed. 

Agentic AI and the hallucination problem 

One element of the partnership’s language is worth attention. IFS Chief Executive Mark Moffat explicitly invoked the hallucination risk in AI systems as a design constraint, not an afterthought. 

“Manufacturers need their factory floor to behave the way it was designed. This partnership with Siemens brings together two companies that each own a critical piece of the puzzle. Agentic AI is the critical frontier, and industrial leaders need solutions with closed loop models and data, and a rich set of context that will not hallucinate in active operations,” Moffat said. “By combining our collective strengths in Industrial AI, we can help manufacturers close the loop between design and reality, and unlock real, measurable performance gains.” 

The concern is not abstract. Agentic AI, systems that take autonomous action rather than simply providing recommendations are beginning to enter industrial environments. In contexts where an AI agent might adjust a production schedule, trigger a maintenance intervention, or modify a configuration, a hallucinated output is not a nuisance, it is a potential safety or compliance event. Designing AI that is grounded in both engineering data and operational history, and auditable across both, is a direct response to that risk. 

Tony Hemmelgarn, President and Chief Executive Officer of Siemens Digital Industries Software, framed the ambition in terms of data infrastructure as much as AI capability. 

“Industrial AI only delivers value when it is grounded in both engineering intent and real-world performance. Together with IFS, we are bringing these domains together by connecting design, manufacturing and asset lifecycle data in a secure, contextualized data fabric. By converging our combined strengths in industrial AI, together we will empower our customers with our vision of an executable Digital Twin that will enable them to accelerate innovation with confidence.” 

A market ready for integration 

The partnership lands at a moment when the industrial software market is consolidating around integration rather than point solutions. The collaboration is intended to improve production planning, maintenance scheduling, and asset management by reducing information silos between engineering, operations, and service organizations. That framing reflects a wider shift in how industrial buyers evaluate technology: not by the capability of individual tools, but by how well those tools connect to the systems already in use. 

Related Articles:

What 500 Business Leaders Got Wrong About Quantum AI in 2025, and What They Think Now

Banks Are Starting to Finance AI Infrastructure Like They Once Financed Airports

AHOY Establishes North American HQ, Partners With Mila Labs, and Launches Canada's First Physical AI Lab