Selon un nouveau rapport de Rockwell Automation, 93 % des fabricants disposent d'un MES, mais seuls 23 % d'entre eux l'ont entièrement intégré.

Rockwell Automation souligne un écart majeur d’exécution du MES: 93% des fabricants en ont un, mais seulement 28% l’ont déployé à l’échelle de l’entreprise et 23% l’ont entièrement intégré avec l’ERP/PLM/contrôle qualité/OT. L’étude pointe aussi que 44% voient l’intégration comme critère d’achat principal, tandis que 46% déclarent avoir subi un incident cybersécurité sur l’année écoulée. Malgré des ambitions IA (42% des processus assistés d’ici un an, 54% d’ici 2030), 43% reconnaissent ne pas utiliser efficacement les données collectées—un frein à la création de valeur.

Analysis

This reads less like a demand inflection for MES and more like evidence that the market has entered the monetization phase of industrial software: the installed base is already there, but wallet share is shifting toward integration, orchestration, and security. That is structurally favorable for platforms with a credible edge-to-cloud architecture and sticky workflow ownership—ROK’s Plex is in the right lane—but it also means the upside is more about higher attach and services mix than a step-change in seat growth. The likely winners over the next 6-18 months are the vendors that can sit between ERP/OT/quality and keep data in a usable state; the losers are point solutions and legacy stacks that force customers into costly middleware projects.

The second-order effect is that AI spend in manufacturing is gated by data hygiene, so near-term budgets should flow first to plumbing, governance, and cybersecurity rather than to flashy AI applications. That favors industrial software, OT security, and system integrators, while hardware-heavy automation names may see slower conversion of interest into orders if customers pause for integration work. The cyber angle is important: more connected plants increase breach visibility, but also lengthen procurement cycles as buyers add compliance reviews and vendor risk checks.

The contrarian read is that this is not a big TAM expansion story; it is a replacement and optimization story. If the market treats the report as evidence of broad AI upside, that is probably overdone until we see measurable improvement in software recurring revenue, retention, and cross-site deployment metrics. The thesis breaks if ROK cannot show Plex-driven bookings or if industrial capex weakens enough that customers defer the integration projects needed to monetize these surveys.

Key Decisions for Investors

• Stay tactically neutral on ROK into the next print; only add on pullbacks if management can show software bookings / recurring revenue acceleration. The report is supportive but not sufficient on its own to re-rate the stock.
• Relative-value: long ROK / short a more hardware-exposed industrial basket (XLI) for 3-6 months if you want to express the view that software integration and OT security will capture a larger share of industrial capex than equipment replacement.
• Watchlist, not a trade yet: Siemens, Schneider Electric, Honeywell, Emerson, PTC, and OT-security names as beneficiaries of the same integration/security spend. Confirm with order growth and software mix before rotating capital.
• Falsifier to the bullish software thesis: if the next 1-2 quarters show flat or decelerating software bookings, or if management emphasizes delayed deployment cycles and lower conversion from MES adoption to enterprise rollout, reduce exposure.
• Avoid buying the AI narrative outright until proof appears in metrics; if ROK’s margins improve without commensurate bookings growth, it may be a services/mix story rather than durable software re-acceleration.