The Built World Mired in 1997 Software (And Why AI is Finally Moving the Needle)
The buildings we inhabit are powered by a brittle, aging software backbone that most people never notice — and that has never bothered to evolve for three decades. The industry that designs, pipes, wires, and anchors our daily lives is massive: roughly $13 trillion in annual construction spend, and yet its core tools remained stubbornly stuck in the late 1990s. Personally, I think this isn’t just a tech hiccup. It’s a structural choke point in a sector that shapes every other aspect of modern life, from housing affordability to data-center capacity for the digital economy.
What is happening here, and why does it matter so much now? The answer sits at three overlapping seams: the monopoly of a single BIM platform, the fragmented, sleeves-rolled-up nature of AEC workflows, and a set of new AI-enabled capabilities finally able to address problems that older software could only pretend to solve. What many people don’t realize is that the real bottleneck isn’t “lack of data” or “not enough clever algorithms” — it’s the mismatch between legacy tools and the velocity and complexity of modern design and build projects. If you take a step back, the opportunity isn’t merely faster drafting. It’s unlocking a capacity revolution in an industry that still commands a huge portion of GDP.
Revit’s grip has become a defining fact of life in AEC
What makes this topic compelling is the way one product has quietly become the system of record for the entire ecosystem. Revit started as a bold ambition in 1997: model a building in 3D and have every change ripple through the design. Autodesk acquired it in 2002 for a modest sum in retrospect, and today Revit commands roughly 95% market share and about $3 billion in annual recurring revenue. In my opinion, that’s less a triumph of product design and more a testament to how hard it is to replace a system that people learn in school, customize for years, and then rely on for mission-critical decisions.
The “taught in schools” effect isn’t just educational trivia; it’s a moat
Almost every engineer or architect enters the workforce already fluent in Revit. Firms standardize on it, clients expect it, and their libraries — custom components, materials, and templates built over countless projects — live inside its ecosystem. That creates a lock-in effect: decades of project history stay siloed in Autodesk formats, making interoperability feel like a perpetual chasing game. What most people miss is how this kills experimentation. If a firm wants to try a new approach, the cost isn’t just a switch; it’s a complete re-education of staff, a migration of libraries, and a risky bet on data fidelity across formats.
Why now, and why AI this time
Two big shifts are tilting the balance. First, large-language models and vision AI now allow software to interpret the rich, messy metadata that lives inside BIM files and the surrounding documents. Room type, equipment specs, code requirements — these aren’t just boxes to tick; they’re drivers of physics, fire protection, cooling loads, and regulatory compliance. AI can infer meaning from unstructured data, semantically classify it, and feed structured inputs to engineering calculations. Second, the industry is staring at accelerating demand for heavy infrastructure and data-center buildouts. The current talent pipeline isn’t keeping pace, so there’s both a practical urgency and a political will to change how work gets done.
Three pathways to transform a $13 trillion market
- Attack Revit directly with a cloud-native, AI-enabled, truly collaborative BIM platform. This is the boldest path and also the steepest climb. Firms have decades of customization and risk averse stakeholders. Yet the moment the problem of “explain to a roomful of tired engineers why a new tool won’t break the model” shifts in favor of demonstrable reliability, a radical shift becomes possible. The strongest signal right now is Motif, the audacious bet to build a fresh BIM stack that can actually rival the trust and depth of Revit. What makes this particularly fascinating is whether an indie challenger can translate AI parity into real-world reliability and training convenience at scale. If they can pull it off, the industry’s inertia might finally crack.
- Build around Revit rather than replace it. The bulk of the actual design work happens outside the BIM authoring environment: Excel spreadsheets, coexisting documents, and a tangle of PDFs. Tools that polish these peripheral workflows can deliver outsized gains without forcing a wholesale re-education. Enter LightTable, which reads construction documents, flags coordination issues, and prioritizes fixes by cost impact. The key here isn’t to upend the system of record but to shrink the lag between design intent and field reality. In my view, this approach is more palatable in the near term and can generate a data-rich feedback loop that makes the eventual replacement easier.
- Target the services layer where the human toll is largest. MEP design, for example, is a sprawling, labor-intensive domain often outsourced and done in ways that scarcely leverage current software. Endra’s AI-native MEP design platform is an emblem of what’s possible when you treat design as a programmable sequence rather than a hand-drawn workflow. This path doesn’t threaten Revit’s dominance so much as it reframes the work pipeline, enabling firms to take on more projects with the same headcount by making repetitive, rules-based tasks automatic. The deeper move here is a capacity expansion: AI unlocks a higher throughput of engineering work and, with it, a new pricing paradigm based on outcomes and efficiency rather than hours billed.
MEP: where the first real crack in the armor appears
The MEP services market isn’t a footnote; it’s a $150 billion growth engine for the broader AEC industry. It’s largely a production line of drawings, calculations, and documentation that must be checked and revised at every step. What’s striking is how little of this work actually happens inside Revit. AI has the potential to turn backlogs into throughput, and that isn’t merely a productivity gain — it’s a structural change in how capacity is allocated across global firms. In this sense, the opportunity feels similar to what we’ve seen in legal or financial services, where rules-based tasks are automated, freeing humans for higher-order problem solving and strategic thinking.
The business model twist that could unlock a megaforce
If you’re looking for a core insight, it’s this: the real value lies not in a better BIM seat but in redefining pricing around outcomes. The industry cannot simply “buy” a faster Revit. It must pay for the incremental margin, reduced risk, and accelerated delivery AI enables. That shifts the conversation from software licensing to performance-based partnerships: a slice of change orders prevented, a portion of time saved, or a percentage point of margin recovered. It’s a shift that makes the opportunity scalable at global scale because capacity is the bottleneck, not the tooling itself.
The bottom line: we are on the cusp of a domain-wide rethink
Today, someone somewhere is waiting for a Revit file to load, or manually performing tasks that feel archaic in a world of intelligent assistants. The waste is real: billions of dollars squandered on rework, delays, and fragile coordination. If we translate the industry’s needs into a new software fabric — one that understands buildings at a structural, systems, and regulatory level, and does so collaboratively — we could unlock a level of productivity that finally matches the scale of demand in housing, transit, and data centers.
So what happens next? Expect three things to intensify: the rise of AI-native BIM platforms that can compete with or complement Revit, a wave of off-the-shelf tools that fix overlooked workflows around the core model, and a service-level revolution where firms pilot performance-based pricing tied to measurable gains. In other words, the built environment is about to get its own economic revolution, driven not by better pixels but by smarter, faster, more trustworthy workflows.
If you’re building in this space, contact a16z’s team to explore partnerships or funding, because the early movers will redefine how civilization’s skeleton is drawn, wired, and cooled for the next generation of growth.
