The Upstream Problem
Lucas opened with the problem cleanly. Structural feedback during early design has always lagged the architectural conversation. The architect tests massing options on Monday. The structural engineer catches up by the following week, often with a single conservative scheme rather than a comparative one. By the time real structural intelligence is available, the building's massing and structural grid have already been set - and the conversation about cost, carbon, and systems integration has happened in the absence of the very analysis that should inform it.
The cost of that lag is not abstract. The biggest structural and embodied-carbon decisions in a building are made in its earliest weeks. The structural engineer who could have shaped them arrives, professionally, a phase too late.
Branch collapses that gap. Real-time analysis and design across materials, with enough depth to be useful rather than indicative. The architect can iterate on massing; the structural response moves at the same speed. Cost, carbon, and systems integration become part of the conversation when the conversation is still open.
Faster Turn-around, Options and Improved Collaboration
As the gap is reduced, it is not simply about time savings but also about providing time to explore options. Should we consider concrete columns and beams? or see if we can achieve a more carbon friendly result with engineered timber?
As these options can be explored in realtime, the architect and the engineer can collaborate and develop a shared understanding and intent. The design is no longer a compromise between engineering and architecture but an aligned, informed and more deeply developed outcome.
Why This Is the Same Story as BIM 2.0
Lucas's talk showed us that Branch is doing for structures what the BIM 2.0 challengers are doing for architecture.
Arcol, Snaptrude, and Motif are, in practice, focused on the early-stage conceptual phase of architectural design. Browser-native, collaborative, fast - built for the part of the design process where decisions move quickly and the cost of being wrong is low. The detailed authoring phase is harder and slower to disrupt; that is why Qonic is the outlier on the architecture side. The challengers are upstream because upstream is where the AI advantage compounds fastest.
Branch is doing the same thing, in the same upstream phase, for structure. Same logic. Same time horizon. Same shift away from authoring-as-bottleneck.
What that produces, if it lands, is a coordinated upstream conversation that the previous generation of BIM tooling could never deliver. Architect, structural engineer, and - eventually - mechanical engineer, all working in something close to real time, against a shared model, during the phase of the project where the building's economics are actually decided. That is not a feature. It is a phase shift.
The Moat Implication
A few weeks ago I wrote that structural analysis was one of the categories where the AEC software moat is weakening fastest. The combination of mature open-source solvers (OpenSees, CalculiX) and AI's ability to assemble interfaces and reasoning around them was compressing the gap between "I have an engine" and "I have a usable product" faster than most product teams appreciate.
Branch is the working example and so much more. It is not a thin UI on an open engine - Lucas was clear that the depth is in their own understanding of structural systems, manufacturing logic, and carbon - but it sits in the part of the market where that combination is now economically tractable.
What that means for the structural analysis incumbents is not a five-alarm fire. Their professional liability and validated datasets are still real. But the assumption that the underlying capability was the moat is no longer safe.
Where This Goes Next
Lucas's roadmap moved downstream into fabrication: more complex geometry, connection design, shop drawings, CNC files. If they execute it, the same upstream tool becomes the input to the supply chain. The structural conversation that began at the massing stage carries through to the steel mill or the timber yard without re-authoring the model into a different system at every handoff.
This is the substrate question Martyn Day and Greg Schleusner have been pushing on, applied to the structural axis. The lifecycle data layer is not just about the operational phase fifty years on. It is also about the design-to-fabrication phase six months on. Branch is one of the first tools I have seen that takes that seriously without pretending the existing handoffs work.
branch3d.com is on my list. If you are an architect who works on anything more structurally interesting than a flat-slab office building, it should be on yours.
The Best Part
While Branch is essentially a startup it sits within the broader context of a Strcutural engineering firm. The problems Branch solves are the problems Lucas and his fellow engineers were facing within their Engineering business. Branch is battle hardened and being built by people who actually use is.
That is a powerful combinatio