For years, manufacturing cybersecurity had a fairly clear boundary: protect the plant.

Secure the industrial network. Restrict access to controllers and engineering workstations. Separate operational technology (OT) from enterprise IT. Monitor critical assets. Keep production running.

That boundary is disappearing.

A modern production environment can depend on an identity service hosted outside the plant, an equipment vendor connecting remotely, an engineering application running in a data center, analytics in the cloud, Industrial Internet of Things (IIoT) gateways moving operational data, and enterprise applications coordinating production and distribution.

Each of those systems may be managed by a different team. From an operational perspective, however, they are part of the same chain.

That creates a fundamental change in the way manufacturers need to think about cybersecurity.

The question is no longer simply: How do we secure the factory? It is becoming: How do we secure everything the factory depends on?

I describe this broader challenge as securing the OT value chain.

The Factory Perimeter Is No Longer the Security Perimeter

Traditional OT security programs understandably concentrated on assets inside manufacturing facilities: programmable logic controllers (PLCs), human-machine interfaces (HMIs), engineering workstations, supervisory systems, historians, industrial servers and the networks connecting them.

Those systems remain critical. But attackers do not have to begin with a PLC to create an operational problem.

Consider a common scenario. A specialist from an equipment manufacturer needs to troubleshoot a production system remotely. The specialist connects through a remote-access environment. Their identity may be managed by one platform, the connection crosses enterprise infrastructure, and the target sits inside an industrial environment.

From the perspective of organizational ownership, these may be several separate systems. From the attacker’s perspective, they are simply a path.

The same principle applies to cloud services, enterprise applications, engineering environments and IIoT infrastructure. A weakness several steps removed from a production asset can still create a route toward operational disruption.

This is why the traditional plant boundary is becoming less useful as the primary definition of OT risk. Manufacturers increasingly need to understand the complete operational dependency chain.

Third-Party Access Shows Why the Model Must Change

Third-party access is one of the clearest examples.

Manufacturing organizations depend heavily on original equipment manufacturers, automation specialists, system integrators and maintenance providers. In many cases, remote support is not optional. When production equipment fails, waiting for a specialist to travel to a facility may significantly extend downtime.

The answer cannot simply be to eliminate remote connectivity. The better question is: How much trust does that connection actually require?

Historically, remote access could mean giving a user network connectivity and then controlling what they could reach through credentials and firewall rules. That approach becomes increasingly difficult to defend in a connected manufacturing environment.

A vendor who needs to troubleshoot one engineering system should not automatically receive visibility into the surrounding OT network.

Instead, access should be tied to a specific identity, approved purpose, target system and period of time. Security teams should understand who is connecting, which asset is required, what privilege is necessary, how long the privilege should remain active, whether the session can be monitored, and whether access can be removed immediately when the work is complete.

These questions turn remote access from a connectivity problem into a governance problem. And that distinction matters.

Zero Trust in OT Cannot Be a Copy of Enterprise IT

Zero Trust provides a useful foundation for addressing this challenge, but applying it to OT requires care.

Industrial environments are not simply enterprise networks with different equipment. Many manufacturing assets have long operational lifecycles. Some cannot support modern security agents. Others depend on legacy protocols or operating systems. Patching can require coordinated production shutdowns. A control that creates a minor inconvenience in an office environment could create a significant operational problem on a production line.

For that reason, Zero Trust in OT should not mean placing authentication controls everywhere and assuming the problem is solved. It should mean reducing unnecessary trust while respecting operational requirements.

That requires understanding identity, device, asset, privilege, communication path and operational consequence together.

Suppose an external engineer requires temporary access to an industrial server. Verifying the engineer’s identity is important, but it is only the beginning. The organization also needs to understand why the access is required, whether the engineer should reach only that server, which protocols are necessary, whether credentials can be protected from direct exposure, how long the connection should remain available and what happens if the session is compromised.

This is where Zero Trust becomes meaningful in OT. The objective is not maximum restriction. It is minimum necessary trust.

Asset Inventory Is Not the Same as Asset Intelligence

The same shift is occurring with asset visibility. Most organizations understand that they need an inventory of their OT assets. But knowing that a device exists is only the first step.

Imagine two industrial Windows systems with the same vulnerability. On a vulnerability report, they may look almost identical. Operationally, they may be completely different.

One may support a non-production test environment with limited connectivity. The other may be an engineering workstation with access to systems supporting a critical manufacturing process. The vulnerability is the same. The risk is not.

That is why OT programs need to move from asset inventory toward asset intelligence.

Security teams need context: what the asset does, where it is located, what process depends on it, what communicates with it, who accesses it, which vulnerabilities affect it, whether its behavior is normal, and what would happen if it became unavailable.

Once those questions can be answered, security teams can make much better decisions about where to focus limited remediation resources.

Stop Counting Vulnerabilities. Start Prioritizing Risk.

This leads to another challenge: vulnerability management.

Traditional vulnerability management tends to be heavily influenced by technical severity. Find the vulnerabilities, rank them and patch the most serious ones. Manufacturing complicates that model.

Patching an OT system may require vendor validation, engineering testing and a scheduled maintenance window. In some environments, an unplanned reboot or application incompatibility can create operational consequences of its own. At the same time, delaying every patch because a system is ‘OT’ is not an acceptable strategy either.

The answer is context.

A practical OT vulnerability decision should consider more than a severity score. It should consider exploitability, exposure, connectivity, asset criticality and the consequence of compromise.

A medium-severity vulnerability on a highly exposed system supporting an important production process may deserve attention before a technically critical vulnerability on an isolated asset with limited operational importance.

And remediation does not always mean immediate patching. Depending on the environment, temporary risk reduction may come from segmentation, firewall restrictions, removal of unnecessary services, tighter access control, application controls or enhanced monitoring until a safe maintenance window is available.

The important measure is not how many vulnerabilities were closed. It is how much meaningful operational risk was reduced.

Segmentation Should Reflect How Manufacturing Actually Works

Network segmentation remains one of the strongest controls available in OT environments. But good segmentation requires more than adding firewalls between networks.

The most useful segmentation discussions begin with operational questions rather than firewall rules: Why do these systems need to communicate? Which protocols are actually required? Who initiates the connection? Should communication be one-way or bidirectional? Does it need to exist continuously? What manufacturing process depends on it? What happens if that connection is compromised?

Those questions often expose unnecessary connectivity that would otherwise remain hidden. They also help security teams avoid implementing controls that look strong architecturally but interfere with legitimate manufacturing processes.

Effective segmentation therefore requires cybersecurity, networking, engineering and operations teams to design the boundaries together. The goal is simple: a compromise in one part of the environment should not automatically provide a path into another.

Cloud and IIoT Have Expanded the OT Trust Boundary

Cloud adoption makes this even more important.

Manufacturers increasingly use cloud platforms for analytics, monitoring, application hosting, predictive maintenance and centralized operational data. IIoT gateways can connect physical processes with applications far beyond the traditional industrial network.

These technologies provide significant operational value. They also create dependencies that did not exist in traditional plant architectures.

A connection from a factory to an approved cloud platform should not automatically be considered low risk because the destination is trusted. Security teams still need to understand what information is moving, how the destination is authenticated, whether communication is encrypted, which identities are involved, whether connectivity is inbound or outbound and what happens to the manufacturing process if the service becomes unavailable.

The important question is no longer simply whether OT is connected to the cloud. It is what trust that connection creates.

Cloud security teams and OT security teams therefore cannot work independently when the same business process crosses both environments.

Technology Alone Will Not Secure the OT Value Chain

Organizations can invest in sophisticated security technology and still struggle with OT risk. The reason is usually not another missing tool. It is inconsistent governance.

Imagine several manufacturing sites handling vendor access differently. One requires approval and time-limited access. Another maintains persistent accounts. A third relies on a locally managed process. The organization may own strong cybersecurity technology, but the actual level of protection varies by site.

The same inconsistency can occur with firewall changes, vulnerability remediation, privileged access, asset ownership and incident response.

This is why governance is essential. Organizations need common principles for secure remote access, identity and privilege management, network connectivity, vulnerability management, third-party security, asset ownership, change management and incident response.

That does not mean every factory must operate identically. Manufacturing sites have different equipment, production processes and operational constraints. Good governance establishes the security outcome that must be achieved while allowing local teams to determine the safest operational method of achieving it.

Cybersecurity and Engineering Need a Common Language

One of the most important lessons from working across IT and OT environments is that cybersecurity decisions become much better when security teams understand operational consequences.

A cybersecurity team may see a vulnerable server. An engineer may see the system required to keep a production process running. Both perspectives are correct. Problems arise when only one perspective drives the decision.

This is why mature OT cybersecurity programs require close cooperation among cybersecurity, engineering, infrastructure, networking, operations, supply chain, cloud teams and site leadership.

Security teams need to understand production. Engineering teams need visibility into cyber risk. Leadership needs both translated into business impact.

The strongest OT security programs are not the ones where cybersecurity controls everything. They are the ones where cybersecurity becomes part of the way operational decisions are made.

The Real Goal Is Operational Resilience

Ultimately, the purpose of OT cybersecurity is not to deploy more security technology. It is to make manufacturing more resilient.

Can the organization identify which operational systems are affected during a cyber incident? Can suspicious third-party access be terminated quickly? Can a compromised environment be contained without unnecessarily disrupting other facilities? Can cybersecurity and engineering teams determine the operational consequence of a vulnerability? Can critical processes be recovered safely? Can production continue when one supporting technology becomes unavailable?

Those questions tell us far more about OT cybersecurity maturity than the number of dashboards, alerts or vulnerabilities an organization has. They also show why securing individual plants is no longer enough.

From Plant Security to Value-Chain Security

Manufacturing has changed.

Production now depends on an interconnected ecosystem of industrial assets, enterprise systems, identities, engineering environments, vendors, networks, cloud platforms and IIoT services.

Every connection creates operational value. Every connection also creates dependency.

The next stage of OT cybersecurity is therefore not simply adding more protection inside the factory. It is understanding those dependencies and applying security across the entire operational value chain.

That means combining asset intelligence, risk-based vulnerability management, precise third-party access, Zero Trust principles, effective segmentation, secure cloud and IIoT integration, and consistent governance.

Most importantly, it means treating cybersecurity as part of operational resilience rather than as a separate technical function.

The factory still matters. But the factory is no longer the boundary.

In modern manufacturing, protecting production means protecting everything production depends on.