Software-Defined Manufacturing Reshapes Industrial Chip Demand

The shift from fixed controllers to programmable infrastructure

Industrial automation is undergoing an architectural transformation that will fundamentally alter which semiconductors manufacturers buy. Software-defined manufacturing (SDM) decouples control applications from dedicated hardware, running virtualized programmable logic controllers and supervisory functions in containers on standard servers or edge infrastructure. This treats automation as programmable infrastructure rather than fixed equipment—a change with significant implications for semiconductor demand.

Traditional automation tightly couples control software to specific PLCs, industrial PCs, or drives. Hardware and software are designed, validated, and replaced together. SDM abstracts applications from hardware through standardized interfaces and virtualization, allowing a virtual PLC to run on an industrial edge server and enabling centralized management without replacing field equipment.

Why it matters

This architectural shift moves semiconductor value away from the large volumes of dedicated industrial microcontrollers, analog components, and single-function devices that have dominated factory floors. Instead, value migrates toward edge processors, heterogeneous system-on-chips combining real-time and general-purpose cores, Time-Sensitive Networking devices, AI accelerators, and hardware security modules. For chip suppliers, this represents not sudden obsolescence but gradual commoditization of standalone controller silicon, with strategic opportunity concentrating in integrated edge compute platforms.

Adoption timeline and hybrid reality

More than half of manufacturers already apply virtualization at the application layer, and a significant share of process manufacturers are evaluating pilots. However, meaningful architectural penetration is expected to take six to seven years even in agile sectors, with broader adoption unlikely before the mid-2030s, according to analysis first reported by Omdia.

Hybrid architectures combining conventional controllers, edge compute, and virtualized applications will dominate the medium term. Physical sensors, actuators, drives, and safety systems remain essential. Real-time control largely stays at the edge or on-premise rather than moving to the cloud. The relevant model is edge-based virtualization with cloud platforms supporting fleet management and data-intensive workloads.

Major automation suppliers including Siemens and Schneider Electric are scaling deployments beyond isolated pilots, demonstrating measurable gains in engineering efficiency, maintenance, energy management, and upgrade speed. Yet interoperability, safety certification, and lifecycle responsibility remain challenging across multi-vendor environments.

The emerging industrial silicon stack

As control functions consolidate onto edge platforms, the industrial semiconductor bill of materials is evolving to include industrial edge processors capable of running virtualized control and AI workloads, heterogeneous SoCs combining general-purpose processors with deterministic real-time cores, industrial Ethernet and TSN networking components, AI accelerators for machine vision and predictive maintenance, higher-capacity memory supporting virtualization, hardware security modules, and functionally safe compute for mixed-criticality systems.

Products from NVIDIA's Jetson platform and Intel's edge processors are appearing in factory specifications that previously belonged exclusively to industrial semiconductor vendors like Texas Instruments, Infineon, STMicroelectronics, and Renesas. Legacy silicon does not disappear—certified physical controllers will continue handling most deterministic, safety-critical loops—but the strategic opportunity lies in heterogeneous SoCs engineered to run hypervisors natively.

Automotive as precedent

The software-defined vehicle transition offers a useful preview. Distributed single-function electronic control units are consolidating into domain controllers and central compute platforms. The result was not the disappearance of automotive MCUs but a more heterogeneous architecture, with new entrants like NVIDIA and Qualcomm capturing the high-compute layer.

Industrial automation is likely to follow a similar pattern, though differences matter: industrial assets often remain in service 15 to 30 years, plants combine multiple equipment generations and vendors, and safety requirements are stringent—all favoring hybrid systems and slower adoption.

Strategic implications

For semiconductor suppliers, the opportunity is not simply supplying more powerful processors but delivering industrial-grade platforms for mixed-criticality environments that combine real-time execution, virtualization support, security isolation, functional safety, and AI capability in single designs. Suppliers viewing industrial demand mainly through the lens of standalone controllers may preserve legacy revenue but risk missing the higher-value compute, connectivity, and security platforms defining the next generation of industrial automation.

These findings were detailed in analysis by Anna Solovieva and Saloni Gankar at Omdia.