What is Industry 4.0?

Industry 4.0 refers to the intelligent, near-real-time networking of machines, products, processes, and IT systems across the entire value chain. Plants exchange data and increasingly make their own decisions in a decentralized way. The physical and digital worlds of production merge.

Technically, Industry 4.0 is the integration of cyber-physical systems into manufacturing and logistics, combined with the Internet of Things. From this emerge the vertical and horizontal integration of value-creation networks, real-time data exchange, and decentralized, self-optimizing decision-making.

The term originated in Germany: it was publicly introduced at Hannover Messe in 2011 and detailed in 2013 with the final report of the Industrie 4.0 Working Group (acatech). Since then the topic has been driven by the Plattform Industrie 4.0. The “4.0” points – like a software version – to the fourth industrial revolution.

Four design principles summarize the idea: interconnection, information transparency, technical assistance for people, and decentralized decisions.

Industry 4.0 is the fourth in a series of technological upheavals. The dates should be read as eras, not exact start dates.

The third revolution is marked by the programmable logic controller (PLC) – the first PLC, the Modicon 084, appeared in 1969. It made individual machines flexibly automatable. Industry 4.0 starts where that automation ends: with end-to-end networking across machine, plant, and company boundaries.

Industry 4.0 is not a single product but the interplay of several technologies. There is no universally binding official list of exactly nine “pillars”; the grouping below captures the building blocks most commonly used in practice and literature.

These building blocks are connected through communication standards. In the field, the PLC supplies the raw data. For connecting to higher-level systems, two standards have become established above all: OPC UA (client/server or PubSub, with information models) and MQTT (lightweight publish/subscribe from edge to platform). Their roles can overlap: MQTT transports messages but does not provide an industrial semantic model by itself.

The automation pyramid

ISA-95 / IEC 62264 describes logical activities and interfaces between enterprise and control domains. The familiar pyramid is a useful illustration, not a prescribed physical architecture. Modern data flows may bridge levels in a controlled way without removing their responsibilities.

The Industrial Internet of Things (IIoT) is the networking of sensors, machines, and plants in industrial settings to collect and exchange operational data. It is the foundation of Industry 4.0: without captured data there is no digital twin, no forecast, and no analysis.

Unlike consumer IoT, IIoT prioritizes reliability, determinism, safety, and very long lifecycles (often 15 to 20 years or more). A failure here is not just inconvenient but can be costly or safety-critical.

The typical data flow:

• Sensors capture readings directly at the process.
• Gateways / edge devices aggregate, filter, and translate protocols.
• Connectivity, for example through OPC UA (structured communication and information models) and MQTT (decoupled publish/subscribe).
• Platform for storage, visualization, and analytics.

A common pattern: an edge gateway acts as an OPC UA client in the field and then publishes the data over MQTT to a cloud broker.

Industrial AI starts with reliable data, not with a model. Readings, timestamps, units, machine state, production order, and quality outcome must be connected before statistics or machine learning can identify a dependable relationship.

• Anomaly detection flags unusual vibration, current, or cycle-time patterns before a fixed limit is crossed.
• Quality prediction relates process parameters to inspection results and exposes causes of scrap.
• Planning and assistance systems support scheduling, energy use, diagnostics, and access to operational knowledge.

Production models need monitoring, versioning, and comparison with a baseline. Safety- or quality-critical decisions still require approval limits, traceability, and a defined fallback path.

A digital twin is a virtual representation of a physical asset, product, or process. Depending on its purpose, it is updated with master data, events, or live data at an appropriate frequency. This makes it possible to simulate, monitor, and analyze behavior and condition.

The concept goes back to Michael Grieves (around 2002). A common distinction is between a product, production, and performance twin. The distinction from a simulation matters: a simulation is a tool that also runs offline. An operational digital twin is linked to its real-world counterpart; required freshness ranges from event-based updates to real-time data depending on the use case.

Example: the digital twin of an injection-molding machine ingests live cycle, temperature, and pressure data to predict part quality and test new parameter sets virtually first.

Cyber-physical systems (CPS) tightly couple physical processes with computation and networking. Embedded computers monitor and control physical processes, often in real-time feedback loops. They create a digital copy of the real world, make decentralized decisions, and communicate with one another over the Internet of Things.

CPS form the conceptual core of Industry 4.0: the fourth revolution is defined precisely by the fusion of physical and digital. IIoT, cloud, and digital twins are the technologies that realize CPS at scale. A networked production made up of such systems is also called a cyber-physical production system (CPPS).

A clear example is an autonomous guided vehicle (AGV) that senses its surroundings, computes a route, coordinates with the MES and other vehicles, and physically moves material – physical action driven by networked computation.

The smart factory is a highly digitized, connected production facility in which machines, systems, and products exchange data to enable largely self-organizing, adaptive manufacturing. It is the integrated outcome in which CPS, IIoT, digital twins, MES, and analytics work together.

It rests on two integration axes:

• Vertical integration: connecting the hierarchy levels within the plant (field → control → MES → ERP) so data flows end to end.
• Horizontal integration: connecting across the value chain – between systems, suppliers, logistics, and customers.

A central promise is lot size 1: the economical production of individually configured products down to a batch of one, through flexible and reconfigurable production.

Predictive maintenance uses condition and sensor data together with analytics and ML models to forecast when a machine is likely to fail. Maintenance is then performed in a targeted way just before failure – instead of on a fixed schedule or only after a breakdown.

The basis is condition monitoring, with continuous capture of vibration, temperature, current, or sound. Vendors and studies report significantly reduced unplanned downtime – but the concrete figures depend heavily on the application and data quality. Predictive maintenance is one of the most easily measurable Industry 4.0 use cases and therefore a typical entry point.

For a deeper look at how predictive and preventive maintenance differ in practice, see our blog article Predictive vs. Preventive Maintenance.

A Manufacturing Execution System (MES) manages and tracks the execution of production in real time – scheduling, quality, traceability, maintenance, and metrics such as OEE. The corresponding manufacturing-operations activities sit at Level 3 of ISA-95, between enterprise planning and process control.

An MES can be an important data hub, but it is not the only possible architecture. Depending on the use case, a SCADA system, historian, unified namespace, or data platform may carry part of the responsibility. Clear semantics, timestamps, quality flags, and data ownership matter more than a product label. Exchange may use OPC UA in OT and APIs, events, or MQTT toward IT. B2MML is MESA International's XML implementation of ISA-95—not the only or prescribed “ISA-95 format.”

Typical flow: the ERP releases a production order, the MES schedules it, dispatches it to the right line, collects OEE and genealogy per unit, and reports completion and quality back.

Industry 4.0 does not end at the factory gate. Suppliers, operators, and customers need selected lifecycle, quality, provenance, carbon, and service data without surrendering control over all raw data. Initiatives under the Manufacturing-X umbrella are developing interoperable, sovereign data spaces for this purpose.

A digital thread connects requirements, engineering, production, use, and service. Asset Administration Shells, shared semantics, identities, and usage policies help partners exchange only authorized information. The digital product passport also depends on this kind of consistent product and lifecycle data.

For the many building blocks to fit together, Industry 4.0 needs common models and standards.

RAMI 4.0

The Reference Architecture Model Industrie 4.0 (RAMI 4.0) is a three-dimensional map that structures Industry 4.0 concepts. It does not prescribe a product but gives experts from mechanical, electrical, software, and IT disciplines a common language to locate where an asset, function, or standard belongs. The three axes are:

• Hierarchy levels (product up to “Connected World”), based on IEC 62264 and IEC 61512.
• Life cycle and value stream (type vs. instance), based on IEC 62890.
• Layers (Business, Functional, Information, Communication, Integration, Asset).

RAMI 4.0 is described in DIN SPEC 91345. The former international pre-standard IEC PAS 63088 was withdrawn on September 27, 2024 and should not be cited as a current IEC standard.

Asset Administration Shell (Verwaltungsschale)

The Asset Administration Shell is a standardized digital representation of an asset and can provide the foundation for a digital twin. It consists of submodels (such as digital nameplate, technical data, operational data) and enables uniform, vendor-neutral access across the entire lifecycle.

It is standardized in the IEC 63278 series (Part 1 on structure was published in 2023) and driven by the Industrial Digital Twin Association (IDTA). Current specifications cover the metamodel, HTTP APIs, AASX package format, registry and discovery, semantics, and security. OPC UA is one mapping and integration option, not the AAS's only transport.

Industry 4.0 deliberately connects the formerly isolated operational technology (OT) – PLCs, SCADA, sensors, machines – to IT networks, cloud, and the internet. Every new connection that makes data usable also expands the attack surface: previously air-gapped plants become reachable, and an incident on the IT side can propagate into physical processes.

Key principles of OT security:

• IEC 62443 is the core standard for the security of industrial automation and control systems.
• Zones and conduits: systems with similar protection needs are grouped into zones; communication between them runs in a controlled way through defined conduits.
• Defense in depth: several overlapping layers of protection instead of a single line of defense.
• Brownfield reality: old, long-lived plants often cannot be patched or taken offline. Protection is then built around the legacy systems, for example through segmentation.

While IT classically puts confidentiality first, OT usually prioritizes availability and safety – because the downtime or manipulation of a physical process can cause immediate harm. For many EU and Austrian operators, the NIS2 directive (EU 2022/2555) is also relevant. Austria has promulgated NISG 2026; its central framework takes effect on October 1, 2026. Whether a company is covered depends especially on sector, size, and role. The EU Data Act has applied since September 12, 2025. Cyber Resilience Act reporting starts September 11, 2026, while its main obligations apply later. This is not legal advice.

Connectivity creates value only when technology, operations, and organization fit together. Common barriers should be visible and owned before the pilot begins.

Not every process needs cloud, AI, or millisecond live data. The smallest architecture that reliably achieves the measurable purpose is often cheaper, safer, and easier to operate.

• Efficiency and productivity: less downtime, optimized throughput, and lower scrap through real-time monitoring.
• Predictive maintenance: failures are forecast; unplanned downtime and maintenance cost fall.
• Flexibility and lot size 1: economical production of individualized products, fast changeovers.
• Transparency and traceability: end-to-end visibility of production, inventory, and quality in real time.
• Resource and energy efficiency: granular monitoring lowers consumption and supports ESG reporting.
• New business models: data-driven services, remote maintenance, and “pay-per-use” instead of pure machine sales.

Industry 4.0 does not have to start big. Especially for SMEs, clearly scoped projects are often the most effective.

• Condition monitoring: retrofitted sensors on a critical machine stream live data to a dashboard – a low-cost, high-impact first step in brownfield plants.
• Predictive maintenance on a press: vibration and temperature sensors detect bearing or tool wear before a stop occurs.
• Digital twin of a line: changes are tested virtually before the physical plant is touched.
• Networked quality data: inline inspection feeds data back to upstream stations; every unit is traceable.
• Energy monitoring: granular metering of machines and lines lowers energy cost – with fast payback.
• Connected products and remote service: machine builders monitor shipped machines in the field and offer remote diagnostics – servitization in practice.

A still smaller starting point is often one useful signal. Our Real Industry 4.0 guide shows how to turn it into a viable use case.

The core difference in one sentence: Industry 3.0 automates individual machines; Industry 4.0 networks the whole – so that data flows and systems partly optimize themselves.

A pragmatic start usually follows the same steps – begin small, create measurable value, then scale.

1. 1

   Assess maturity and define the problem

   Review assets, information systems, organization, and skills, then select one concrete operational problem.

2. 2

   Establish the baseline

   Record a reliable KPI baseline such as downtime, scrap, OEE, energy per unit, or changeover time.

3. 3

   Audit assets, data, and risk

   Review interfaces, signal quality, timestamps, semantics, network design, safety, and OT security together.

4. 4

   Define the pilot and ownership

   Set a bounded use case, accountable owner, acceptance criteria, budget, and rollback plan.

5. 5

   Connect read-only and validate

   Prefer connections that leave control logic untouched, then validate production stability and data quality.

6. 6

   Measure impact and lifecycle cost

   Compare the result with the baseline and include operations, maintenance, licenses, and training in the payback calculation.

7. 7

   Standardize and scale

   Document proven data models, security rules, monitoring, and operational ownership before adding more assets.

The acatech Industrie 4.0 Maturity Index helps assess capabilities across resources, information systems, organization, and culture.

Primary sources for the definitions, standards, and regulatory status in this article (accessed August 17, 2026):

• acatech: Recommendations for implementing Industrie 4.0 (2013)
• acatech: Industrie 4.0 Maturity Index
• ISA: ISA-95 / IEC 62264 Enterprise-Control Integration
• ISA: ISA/IEC 62443 for industrial cybersecurity
• IEC: IEC 63278-1 – Asset Administration Shell
• IDTA: current AAS specifications
• IEC: IEC PAS 63088 (withdrawn)
• RIS: Austrian NISG 2026
• European Commission: Data Act
• European Commission: Cyber Resilience Act
• European Commission: Industry 5.0