Aize has acquired French industrial AI company Samp, combining two software approaches intended to give engineering and operations teams a more reliable representation of complex industrial assets. The deal brings Samp’s Shared Reality technology into Aize’s industrial data platform, linking information captured from physical facilities with engineering, operational, and enterprise data across both greenfield developments and ageing brownfield plants.
The acquisition addresses a persistent problem in process and infrastructure industries: the plant that exists in the field does not always match the plant described in the documentation. New facilities may begin life with structured CAD models and comprehensive engineering data, but older assets accumulate replacement equipment, modifications, redlined drawings, isolated databases, and undocumented changes over decades of operation. The gap becomes expensive whenever maintenance, inspection, shutdown, or project work depends on knowing what is actually installed.
Aize approaches that problem from the engineering-data side. Its platform brings together 3D models, engineering information, operational data, and enterprise systems in a shared environment. Samp starts closer to physical reality, using data from laser scanning and other reality-capture methods to construct an interactive representation of the current asset and connect it with tags, drawings, piping and instrumentation diagrams, and associated engineering information.
The distinction matters particularly in brownfield facilities because recreating a complete CAD model can become a substantial engineering project in its own right. Samp’s Shared Reality platform is designed to identify equipment and piping within captured data and turn them into structured, navigable assets without first requiring a conventional remodelling exercise. That provides a route to comparing the actual installation with the drawings and databases on which maintenance and modification work would otherwise rely.
Samp’s technology has been used across more than 500 industrial facilities, particularly in energy and water, while Aize already works with operators and engineering businesses responsible for large industrial assets. The combination therefore begins with an operating installed base rather than two experimental technologies searching for their first application. Both companies have concentrated on environments where inaccurate asset information can translate directly into additional engineering hours, site visits, shutdown risk, and maintenance cost.
The immediate engineering value is likely to be in work preparation. A brownfield modification can begin with hours of document searches, drawing checks, walkdowns, measurements, and verification before the design team is confident that the information supplied represents the plant. Better alignment between scans, P&IDs, equipment tags, and operational systems can reduce that uncertainty and give contractors, operators, and remote engineering teams a common starting point.
Shutdown planning provides another obvious application. Maintenance and modification scopes are often prepared months before access is available, and discovering a documentation discrepancy once equipment has been isolated can consume an expensive outage window. A reality-linked engineering model cannot eliminate unforeseen conditions, but it can move part of the verification work earlier and make visible differences between documented and installed equipment before crews reach the job.
The transaction also reflects a wider change in the way industrial digital twins are being developed. Early implementations were frequently centred on accurate design models, but an immaculate representation of the original engineering has limited operational value if years of physical changes have made the facility different. A current scan presents the opposite problem: it captures geometry but does not automatically contain equipment duty, process relationships, operating history, or maintenance context.
Combining those two information sets gives Aize a broader proposition, although integration will determine whether it works outside demonstrations. Industrial operators already maintain document systems, historians, maintenance platforms, engineering databases, and specialist applications that cannot simply be discarded following a software acquisition. Aize has emphasised an open ecosystem approach and intends to continue working with specialist reality-capture providers rather than trying to own the entire scanning process itself.
That interoperability is important because another closed information repository would simply add to the problem the acquisition is intended to address. The useful role for a digital twin is as a working layer across information that already exists, helping engineers move between a physical asset, its engineering definition, and current operating data without spending the first part of every task establishing which source can be trusted.
Aize plans general availability of its next-generation product before the end of 2026, with work intended to improve performance and movement between 3D assets, engineering drawings, and industrial information. Financial terms for the Samp acquisition were not disclosed, so the commercial scale of the transaction remains unknown.
The industrial software market has no shortage of expansive claims for digital twins. The more immediate problem is less glamorous and considerably easier to measure: whether an engineer can find dependable information before somebody alters a live plant. Bringing Samp’s view of physical reality into Aize’s engineering environment gives the combined business a credible route into that problem. The harder task will be maintaining that accuracy after the next modification is completed.
