For twenty years, the construction industry’s dirty secret has been that its most expensive software products don’t talk to each other. A structural engineer models in one platform, the architect works in another, the general contractor coordinates in a third, and somewhere between design and delivery, data gets exported, flattened, re-imported, and quietly corrupted. The workarounds became their own cottage industry: middleware vendors, conversion consultants, and entire staff positions dedicated to shuffling files between systems. But over the past eighteen months, something has genuinely shifted. The vendors who spent decades building walled gardens are opening gates, the standards bodies are rethinking their foundations, and a new generation of platforms is treating interoperability as the product rather than an afterthought. Here is where the design-to-delivery data pipeline actually stands in 2026, and why this round of promises looks different from the ones that came before.
The Autodesk-Nemetschek Agreement Is Producing Real Features
When Autodesk and the Nemetschek Group announced their interoperability agreement in April 2024, the industry reaction was understandable skepticism. Two competitors agreeing to open their APIs to each other sounded like a press release, not a roadmap. Two years later, the shipping features tell a different story.
The most concrete example surfaced at AIA 2026 in June: Archicad users can now upload a 3D view or a full model directly as an Autodesk Forma Data Exchange, making the geometry immediately accessible as reference data inside Revit. Graphisoft has added Publisher Set syncing between Archicad and Autodesk’s cloud data management environment, which means an architecture firm running Archicad can participate in a Forma-based project without the export-import gymnastics that used to define cross-vendor collaboration.
This matters beyond the specific features. Architecture and engineering firms rarely get to dictate what software their partners use. A workflow where an Archicad model flows into a Revit-centric coordination environment without a lossy IFC round-trip removes one of the most persistent friction points in multi-firm project delivery. Nemetschek, for its part, is reportedly building its own cloud-based early-design ecosystem targeted at 2027, which suggests the two companies see interoperability as a competitive necessity rather than a charitable gesture.
IFC 5 Is a Rethink, Not a Revision
The Industry Foundation Classes standard has been the backbone of vendor-neutral BIM exchange since the 1990s, and it has carried that weight with a data model that shows its age. IFC’s file-based, monolithic structure was designed for an era of desktop software and complete-model handoffs. It was never built for granular, cloud-native, query-what-you-need workflows.
buildingSMART’s answer is IFC 5, and the early technical direction is a genuine departure. The alpha version released for community feedback restructures the standard around composable, layered data rather than monolithic schema files, borrowing concepts that will look familiar to anyone who has watched the OpenUSD ecosystem develop. Development is happening in the open on GitHub, with active issues and community feedback shaping the specification through 2026.
The honest caveat: IFC 5 is at alpha stage, and standards bodies move at standards-body speed. Nobody should be making 2026 project decisions based on IFC 5. But the direction matters, because it signals that the industry’s canonical open standard is being rebuilt for the way data actually moves now: incrementally, via API, between cloud services, rather than as enormous monolithic files attached to a transmittal.
OpenUSD Is Entering Construction Through the Digital Twin Door
While buildingSMART rebuilds IFC, another open standard is arriving from an unexpected direction. OpenUSD, the scene-description framework originally developed at Pixar, hit a major milestone when the Alliance for OpenUSD published its Core Specification 1.0 in December 2025. The alliance is now advancing Core Specification 1.1 through 2026 and has begun the ISO certification process, the first step toward the kind of international standardization that government and enterprise procurement teams require.
For construction, the interesting question is where USD fits alongside IFC rather than whether it replaces it. IFC carries the semantic richness of building elements: this is a wall, it has a fire rating, it belongs to this level. USD excels at aggregating massive, heterogeneous 3D data from many sources into one unified, queryable environment. Trimble has been vocal about OpenUSD’s potential to address BIM interoperability gaps in digital twin and collaboration workflows, aggregating CAD, geospatial, and IoT sensor data into a single environment. The plausible outcome is a layered future: IFC for building semantics, USD for aggregation and visualization at scale, with the two standards increasingly designed to interoperate rather than compete.
The Connector-less Platforms Are Winning Converts
Perhaps the clearest evidence that interoperability is becoming a product category in its own right: the money is following it. In April 2026, Suffolk Technologies, the venture platform affiliated with Suffolk Construction, invested in Speckle, the open-source data platform that extracts and normalizes BIM data from tools like Revit, Civil 3D, Rhino, Archicad, and Tekla into a unified, queryable format.
Speckle’s recent product direction is worth watching closely. The platform now supports direct Revit file imports without requiring a Revit connector or license, and as of April 2026 it can ingest Navisworks models directly or sync them from Autodesk Construction Cloud. That is a meaningful shift: instead of asking every project stakeholder to install and maintain plugins, the platform meets the files where they are.
The strategic driver behind the Suffolk investment is telling. The stated goal is unlocking the enormous share of captured project data that currently goes unused and making it available for AI workflows at scale. That framing has quietly become the strongest business case for interoperability in years. Machine learning models are only as good as the data they can reach, and data locked inside proprietary authoring formats is unreachable. Every contractor building an AI strategy is discovering that their interoperability problem and their AI problem are the same problem.
What This Means for the Market
Step back from the individual announcements and a pattern emerges. The forces that historically kept construction data siloed, chiefly vendor lock-in economics and the technical inertia of file-based exchange, are both weakening at the same time. Vendors are opening up because their largest customers are demanding it and because cloud platform economics reward ecosystems over walled gardens. The standards are modernizing because cloud-native, granular data exchange has become the baseline expectation rather than an aspiration.
The AI wave is accelerating all of it. Owners and contractors evaluating AI tools are asking a question that never used to appear in software procurement: can this system access all of our project data, regardless of which tool authored it? That question rewards open platforms and punishes closed ones, and the vendors can read the market as well as anyone.
None of this means the interoperability problem is solved. Data still degrades across tool boundaries, contractual and liability questions around shared models remain genuinely hard, and plenty of project teams are still emailing files because that is what the specification requires. But the gap between design data and delivery data is narrower today than at any point in the BIM era, and for the first time, the largest vendors have aligned commercial incentives to keep narrowing it.
Where Do You Stand?
If you are running projects today, the practical question is whether your data pipeline is positioned to benefit from any of this. Are you still rebuilding models at phase handoffs? Is your coordination data accessible to anything beyond the tool that created it? I would like to hear how interoperability is playing out on real projects: what is working, what still breaks, and which vendor promises have actually materialized on your jobsites. Share your experience in the comments, or reach out directly on LinkedIn — the best material on this site starts as a conversation with someone who has fought these battles on a live project.