What's Changing on Your R&D Floor

Every CTO at a mid-size European industrial automation vendor has had the same Monday. Your VP Firmware walks through the roadmap. New drive family in Q4. Safety module revision in Q3. CODESYS-based application-engineering library expansion all year. Your CFO forwarded a Siemens investor update on the way in: "TIA Copilot in production at Audi and Bosch mid-size divisions; sixteen-percent TIA Portal productivity lift." Your VP Product mentions the CDO's Industry 4.0 Q2 budget ask came in at €4.8M for a horizontal AI rollout. Your Head of Service Platform was in your office Friday asking for a ringfenced €1.2M for a service-diagnostic data graph across the top fifty installed-base customers. You told him it might have to wait a quarter.

You have run this function for fifteen years. The pattern used to be simple. Hire strong firmware architects. Hold the IEC 61508 line. Ship the next drive generation. Meet the board on firmware release velocity and silicon headcount. The business was understandable.

Something changed. What used to take fifty control engineers now runs on five, and the same compression is landing on application-engineering tooling at the platform clock. Siemens TIA Copilot is production in 2025. Schneider EcoStruxure AI is production in 2025. Rockwell, ABB, Mitsubishi are each within twelve months. Your application-engineering depth — the library reusability, the commissioning automation, the HMI generation — used to be one of your differentiators. That layer is inside the absorption path of every big-vendor platform your mid-size customers run on.

You are not running one R&D function. You are running five engineering groups on one hurdle rate, and three of them sit inside the Siemens and Schneider absorption surface. Firmware and silicon are narrowly inside. Control software is inside. Application-engineering tooling is deeply inside. Your service-diagnostic platform is outside, because it runs on your installed base, which the absorbing platforms do not have. Your integration-and-commissioning work sits half inside, half outside. Fund them as one pool and the absorption clock eats the pool. Fund them as two budgets with two hurdle rates and the compounding layer survives.

This briefing covers what is happening in your sector, what it means for the R&D function you run, and the three questions your team should answer before the next Q2 board cycle.

Siemens TIA Copilot, built with Azure AI, is in production release during 2025. Schneider Electric's PLC Code Generation Copilot inside EcoStruxure reduces development time thirty to fifty percent. Rockwell, ABB, Mitsubishi are each shipping their versions within twelve months. Every mid-size customer on a Siemens, Schneider, Rockwell, ABB, or Mitsubishi stack will have AI sitting inside their own application-engineering environment by the end of 2026. The tooling you spent fifteen years differentiating on — library reusability, HMI generation speed, commissioning script quality, documentation depth — is what those platforms are absorbing first. Your depth on those layers was real. Your depth on those layers was also exactly what the big vendors shipped the AI into.

The moat your firm ran on for a decade has moved. The application-engineering layer commoditises inside the absorbing platform over the next twenty-four months. A capability that used to take your 40-engineer application team three years to build can be rebuilt by the platform's AI in twelve to eighteen months, delivered to the customer inside the TIA Portal or EcoStruxure workspace they already run. Your feature edge on application tooling was never going to hold against a big vendor's platform absorption. What can hold is the layer underneath the model and the layer beside it: what requires silicon iteration, IEC 61508 and IEC 62443 compliance, vision algorithms trained on your imaging hardware, motion-control loops tuned against twenty years of customer edge cases. And the data layer the absorbing platforms cannot replicate: the telemetry, operating data, and maintenance histories from your installed base. Those compound. They take years to build and they cannot be bought.

Your application-engineering feature shipped in 2024 is on an eighteen-month absorption curve. Your service-diagnostic data graph across your installed base is the asset Siemens and Schneider cannot replicate without your customers.

The sector's speed mismatch is the third force. Siemens ships platform-level AI on a quarterly cadence now. Schneider ships platform-level AI on a quarterly cadence now. Your firm ships a product generation every twenty-four to thirty-six months and your CDO's Industry 4.0 roadmap was approved in 2023. You cannot pivot at platform speed on platform terms. What you can do is identify which of your engineering groups sits outside the absorption surface, stop defending the groups that sit inside it, and ship the moat the absorbing platforms cannot. The firms already moving are the ones where the CTO walked into Q1 with two R&D budgets, not one. The ones still running a horizontal AI rollout across all five engineering groups are funding Siemens's absorption from inside their own P&L.

Your build engine. Your control engineers are shipping more application-engineering output per week than they have in a decade. Your safety-rated firmware team holds the IEC 61508 line and cannot take the same tool-layer compression. Your VP Firmware cannot tell you whether the review bar that worked in 2020 holds against agent-generated control code. Ninety-five percent of corporate AI deployments produce no measurable P&L impact, and the gap between what the tools ship and what your build engine actually absorbs is widening every quarter the review discipline is not redesigned.

Your platform and data. The absorbing platforms — Siemens TIA Copilot, Schneider EcoStruxure AI, Rockwell FactoryTalk — are where the tool-layer AI lives. You cannot win that fight on their ground. You can buy the build-side discipline from the tool layer where your customers accept it, and redirect the compounding capital into the layer the absorbing platforms cannot follow: the service-diagnostic data graph built from your installed base's telemetry, data rights renegotiated across the next twenty-four months of customer renewals, integration depth published as structured metadata for your top fifty customer installations. A Siemens TIA Copilot cannot replicate that. Siemens does not have your customers' specific operating data.

Your product thesis. Your board asked for the firm's three-sentence thesis last quarter. Your answer was the one from the 2023 strategy offsite. Your rate card assumes the application-engineering layer is still a margin-bearing differentiator. Siemens ships AI into it inside their own platform now. That is a capital-allocation question about which of your engineering groups sits outside the absorption surface, not a pricing question you can solve in a commercial meeting. Your CFO and commercial leader need a joint document with you before the first outcomes-priced service contract lands from an absorbing platform inside your installed base.

Your R&D bench. Your senior firmware architects carry the debug instinct behind every motion-control library your firm ships. Two of them retire in 2028. Your senior vision engineer carries twenty years of edge-case judgment on your sensor portfolio. The embedded-systems junior pathway at European universities has thinned every year since 2018. No routine work, no training ground, no senior bench in 2030. OpenAI and Anthropic resumed junior hiring in late 2025 specifically to pair juniors with seniors on judgment-intensive work. The firm that cuts application-engineering headcount for this year's margin has a cheaper function in Q4 and no mid-tier bench for the next product generation.

Your board-level authority. Thirty percent of European AI-strategy ownership sits with the CIO, twenty-three percent CDAO, twenty-one percent CAIO, one percent COO. You are not on the list. Your CDO was hired in 2023 to run an Industry 4.0 roadmap that predates the TIA Copilot release curve. Every month without a written CTO-CDO-CEO boundary, the compounding-capital question drifts further from the seat that can defend it. Your Q1 boundary agreement lands before the next CDO budget request, not after.

Your sector is repricing from the platform down because Siemens TIA Copilot and Schneider EcoStruxure AI are absorbing what used to be your application-engineering differentiation. Your function is compressing because what used to take fifty control engineers now runs on five, and the review bar that worked for fifty cannot hold the volume five produce with agents. Same force, two altitudes. There is a third thing happening that most mid-market industrial CTOs have not named yet.

The moat that matters in 2029 is not inside the application layer. It sits in silicon below it and in the service-diagnostic data graph beside it. The application feature you ship this quarter is on an eighteen-month absorption curve. What does not commoditise: the silicon you iterate, the vision algorithms trained on your imaging hardware, the motion-control loops tuned against twenty years of customer edge cases, the safety-rated firmware, and the operating data from your installed base. A big-vendor platform cannot replicate that layer at platform scale because they do not have your customers' plants. You do. The window closes at the rate Siemens and Schneider ship their AI into your mid-size customers' commissioning cycles.

In the late 1990s a generation of independent software-layer players defined the sector. Wonderware on HMI and SCADA. Intellution. Citect. By 2015 every one had been absorbed into the big vendors or was competing inside their platforms. The mechanism was not that independents shipped worse software. It was that the big vendors' platforms absorbed the software layer at a price independents could not match. A handful survived by moving up the data-asset curve. The 2026 version is the same mechanism at a faster clock.

You have eighteen to twenty-four months. The CTOs who claim the authorship of this decision in Q1 author the decade. The ones who wait inherit a seat narrowed to firmware uptime and a scorecard that stopped mattering the quarter TIA Copilot shipped into their biggest customer's plant.

Every industrial automation CTO has had the AI conversation. Most are making the same four decisions. A smaller group is making different ones.

Three questions every industrial automation CTO is tracking. The third is the crux. The first two are how you earn the right to answer it.

Four concrete moves any industrial automation CTO can start this quarter. Each addresses one sub-function.

The mid-market industrial automation CTO has four places to go. None can do what we do.

The big industrial-transformation practices ship standardised Industry 4.0 frameworks scaled across a sector book. Cost structure requires reusability. Your seat is idiosyncratic against specific absorbing platforms. Governance theatre the board will not fund twice.

The hyperscaler vertical-industries teams price the engagement as enablement for cloud consumption. Every build-or-buy decision routes through their pricing for three years.

The big-vendor software-platform arms have a quote on your desk. Siemens, Schneider, AVEVA, Rockwell. They are the absorbing platforms your moat sits against. They cannot honestly advise on which of your layers to harden against themselves.

An in-house or fractional CDO or CAIO splits the seat. Two-thirds of newly-appointed CAIOs spend year one compiling AI inventories. Your 2023 CDO runs an Industry 4.0 roadmap that predates the TIA Copilot release curve.

The four options you have already looked at. None can run the capital-allocation conversation the seat requires. GRAIL is the fifth.

We start with your R&D leadership team because the change cannot outrun their personal understanding of what is possible. Three workshops. Each delivers standalone value. Each earns the right to the next.

The questions where your R&D leadership cannot agree on an answer are the ones worth an hour on the agenda. Ask these honestly.