The future of virtual PLCs, 2026 to 2030

Nobody can date the next five years of industrial control, so this article separates what has shipped, what is a standard in progress and what is our own judgement. Each product claim is tied to a source listed at the end, and where we could not confirm a current status, we say so.

For the basics of what a virtual PLC is, read virtual PLCs: control logic without the controller. For the mechanism, read how a virtual PLC works.

Vendor virtual controllers are products now

Two vendors illustrate the shift. Siemens announced the SIMATIC S7-1500V on 24 May 2023 as an entirely virtual PLC based on the S7-1500, independent of hardware. It runs on computers with the Industrial Edge runtime, it is described as completely compatible with TIA Portal so existing projects can be reused, and updates are managed centrally through Industrial Edge Management.

CODESYS sells Virtual Control SL, and the Intel page for its embedded world 2026 demo shows multiple instances in Docker containers on Linux, on an OnLogic HX521 with Intel Core Ultra processors. The demo took place on 10 to 12 March 2026 in Nuremberg.

What this means for 2026 to 2030, in our judgement: the first question about a new line will be whether the logic can run virtually, and the second will be what the runtime depends on. The attraction is that logic can be versioned, backed up and moved like software. The cost is that the platform becomes an IT asset with an IT lifecycle: operating system, hypervisor, patches and a server to run them.

Containers and the platform underneath

A container is a way to ship a runtime with its dependencies. The CODESYS safety data sheet lists its delivery form as a container image based on the OSADL base image. OSADL is the Open Source Automation Development Lab, which publishes a real-time Linux base.

Containerised control brings the habits of software delivery to automation: images, registries, orchestration, rollback. It also brings a new question every plant has to answer: who owns the Linux patch cycle. Expect the work to move toward platform engineering, and expect controls engineers to be asked for it.

IEC 61499 and portability

IEC 61499 is an international standard, first published in 2005, for distributed control systems. Where IEC 61131-3 runs cyclically, IEC 61499 is event-driven: function blocks have event inputs and outputs as well as data, and an application is a network of function blocks mapped onto devices afterwards. Part 2 defines an XML exchange format so applications can move between vendors' tools.

Eclipse 4diac is the best-known open source implementation. It has an IDE and a real-time capable runtime called FORTE. Its own news page lists weekly updates in October 2026, including OPC UA node creation and event handling improvements in FORTE, so it is an active project.

Our judgement is that IEC 61499 will remain a smaller part of the market than IEC 61131-3 through 2030, with the most interest where distribution and vendor independence matter most. The practical consequence for a PLC programmer is small: IEC 61499 builds on the same function block ideas.

O-PAS and open process automation

The Open Process Automation Forum, part of The Open Group, published the O-PAS Version 2.1 Preliminary Standard on 27 May 2021. Its announcement describes a reference architecture and information model so that a distributed, heterogeneous ecosystem can interoperate, with the stated aims of interoperability and portability, lower lifecycle cost and more freedom in managing obsolescence. It said a final 2.1 was scheduled for the first quarter of 2022, with a Version 3.0 to follow that would further detail the physical distributed control platform.

We could not retrieve the current standard when writing, so we do not state which version is current or how many products conform. Check The Open Group's O-PAS pages for that. The direction is clear, though: process plants asking for control that is not tied to one vendor's hardware.

Time-sensitive networking

Time-sensitive networking is a set of IEEE 802.1 standards for sending time-critical data over Ethernet. IEEE 802.1AS gives devices a shared clock. IEEE 802.1Qbv, published in 2016 as 802.1Qbv-2015, adds a Time-Aware Shaper that divides traffic into repeating time cycles. Together they can provide bounded latency, and a profile for industrial automation, IEC/IEEE 60802, is in development according to Wikipedia.

The same source says that deriving guaranteed delay bounds in TSN is still an active research topic. It is a reason to be careful with claims about a plant network that has "TSN" printed on a datasheet. For a virtual PLC, deterministic networking is the missing piece between a controller in a rack in the server room and an I/O device at the machine.

Virtual commissioning will get more routine

Virtual commissioning means testing control logic against a model of the machine before the machine exists. Automation World describes a Siemens-based example at Burr Oak Tool: the machine's CAD model was animated, driven by the actual PLC code running on PLCSIM Advanced, with identical tag names on the virtual and the real system. The article reports a build shortened by four to six weeks and says the PLC code needed no change moving to the real controller. It is a single user case and the claims come from the vendor and one user, so treat the numbers as an example, not a benchmark.

It is cheap to try, and it fits CI. We expect more teams to run the same logic against a model on every change. See virtual commissioning explained and digital twins for PLC training.

AI-assisted programming has to be compiled and tested

Large language models will write more PLC code between now and 2030. The question for control engineering is not whether, but what stands between a draft and a machine. Our view: nothing generated should be trusted until it has been compiled, run against a model that can fail it, and reviewed by someone accountable. A virtual controller is the natural place for the middle step. See AI PLC code: compile and run and why ChatGPT ladder logic fails.

Questions to ask before you adopt one

A virtual controller changes who is responsible for what, so the questions are about the platform as much as the logic.

• What is the timing guarantee, and who measured it? Ask for the platform, the kernel and the load it was measured under, not a typical figure.
• Who patches the host? The controller image and the operating system underneath it age at different rates.
• What happens when the host fails? Find out whether a second instance takes over, how long that takes and what the outputs do meanwhile.
• How is the network scheduled? If the answer is "it is Ethernet", the control loop shares a wire with everything else.
• Can the logic leave? IEC 61131-3 code moves between tools with effort. Ask what is lost in the move, and whether an open exchange format is supported.
• What is certified, exactly? A certificate names a product, a version and conditions. Read them.

None of these is new. Each was asked of PLC hardware once, and the answers sat on a data sheet. For a virtual controller, the answers sit across several documents from several suppliers, which is why they are worth asking early.

What will not change

• Hard real time at the edge. A loop that closes in a millisecond on a drive or a servo wants a deterministic target. Virtual controllers close the gap with real-time kernels, and TSN with a deterministic network, but both have to be specified, not assumed.
• Evidence for anything that protects people. A virtual safety controller can be certified. The CODESYS data sheet lists a SIL 3 product certificate issued by TÜV Süd, a maximum of 1 MB of application code and 1 MB of data, and a safety response time of two times the watchdog time plus two times the safety task cycle time. That is a specific product with stated conditions. It does not make virtual PLCs safe in general, and it does not apply to a simulator.
• Lifecycle discipline. Hardware gets obsolete. A virtual runtime gets patched. Both need an owner.
• The fundamentals. The scan, the image, the timer and the state machine are what the virtual controller runs too.

What this means if you are learning PLCs now

1. Learn the fundamentals first: the scan cycle, timers, edges, and state machines.
2. Learn Structured Text alongside ladder, since it is how most larger programs are written and how generated code arrives.
3. Get comfortable testing logic against a model of the process, not only against yourself.
4. Learn enough Linux, containers and networking that a platform is not a mystery.
5. Keep a healthy suspicion about claims. Read the data sheet, find the conditions.

Where LADV fits

The Wartens LADV virtual PLCs are a learning and test tool. They run in a browser, solve ladder and Structured Text scan by scan and give you a process to test against. They are not a production runtime, not real time, not certified and not safety rated, and they have no fieldbus. Their use is to practise the habits above, free, with nothing to install: open the lab, read the series comparison or follow the hands-on guide. For hands-on 3D machine scenes and guided courses, EDWartens Lab (https://lab.edwartens.com) is one place to practise.

How we tested

On 2026-11-01 we read the sources listed below and checked each product claim against them. The Siemens, CODESYS, Intel, Open Group, Eclipse 4diac, Wikipedia and Automation World pages were fetched. The CODESYS Virtual Control SL product page and the current O-PAS standard could not be fetched, so we make no claims drawn from them. This page contains no Structured Text program. Predictions are our judgement and are labelled as such. We did not test any vendor product.