Everything Is Going to Change — Martyn Day on AEC's AI Reckoning

Why It Matters

The shift erodes the core revenue model of AEC firms, accelerating a move toward data‑centric, value‑based services that could reshape industry economics and competitive dynamics.

Key Takeaways

• MEP solvers generate full layouts overnight from Revit files
• Real-time FEA instantly updates structural analysis with design changes
• AI replicates ~70% of CDE functions, reducing manual tasks
• Open SDKs erode software moats, exposing CDEs, clash detection
• Data ownership and cleaning give firms competitive advantage over vendors

Pulse Analysis

The AEC sector is at a tipping point as AI‑powered automation compresses months of engineering coordination into hours. MEP solvers now produce complete electrical and mechanical schematics overnight, and real‑time finite‑element analysis reacts instantly to design tweaks. This speed eliminates the traditional justification for billable‑hour pricing, prompting firms to explore value‑based billing models that align revenue with outcomes rather than time spent. Early adopters that integrate these tools can deliver projects faster, reduce labor costs, and win competitive bids, while laggards risk obsolescence.

Simultaneously, the rise of open software development kits (SDKs) and data‑layer platforms is dismantling long‑standing software moats. Vendors like Palantir are positioning themselves not merely as tool providers but as custodians of the data that fuels every downstream workflow. This shift threatens the profitability of named‑user licenses and token‑based pricing, as firms increasingly view software as a conduit for their own proprietary data. Companies that own, cleanse, and train on their project data will command a strategic advantage, leveraging insights that off‑the‑shelf solutions cannot replicate.

Beyond technology, the workforce challenge looms large. The UK’s apprenticeship collapse signals a looming talent shortage, while multi‑agent AI systems promise to automate complex decision‑making tasks. Firms that reshore data, invest in internal AI capabilities, and restructure governance to accommodate autonomous agents will be better positioned for the next wave of disruption. In the coming 12‑24 months, AEC leaders must align technology, data strategy, and talent development to stay competitive in an increasingly AI‑driven marketplace.

Original Description

The architecture business model was built for a world where hours spent on the project equals value. AI doesn't work that way.

--------

In this episode, I sit down with Martyn Day, journalist and publisher at X3DMedia — the company behind AEC Magazine and Develop3D — to dig into what it means when engineering automation, open SDKs, and Palantir-scale data plays arrive in AEC at the same time. Martyn has covered this industry for 37 years. He says this moment is categorically different from everything before it.

MEP solvers can generate a complete electrical and mechanical layout overnight from a Revit file. Real-time FEA tools let structural analysis respond instantly to a floor plate move. AI can already replicate 70% of what a CDE does. And companies like Palantir aren't arriving to sell software — they're arriving to own the data layer between every downstream process, making everyone else just a pipe feeding their system.

The billable hour model is in the crosshairs. So is the named user license. Martyn calls it the asteroid: the technologies are already visible on the horizon; the impact is still ahead of us.

This conversation is for firm leaders, technology directors, and BIM managers who sense the ground shifting and need to understand what it means for their practice in the next 12 to 24 months.

What you'll learn in this episode:

• Why MEP solvers are compressing months of engineering coordination into hours — and what that does to how architects and engineers collaborate

• How open SDKs are dismantling software moats, and which products (CDEs, clash detection, model checking) are most at risk

• Why firms that own their data, clean it, and train on it will have advantages no vendor can replicate

• What the UK engineering sector's apprenticeship collapse tells us about AEC's workforce future

• Why we're at peak named user licensing, and how layered cloud and token costs are changing firm economics

• How multi-agent systems represent the next wave — and why most firms aren't remotely prepared

Chapters:

00:00 Introduction

00:02:50 How Martyn tracks AEC trends with AI agents

00:07:00 Engineering automation: MEP solvers and real-time FEA

00:10:20 AI and the architecture billable hour model

00:11:30 Palantir's AEC play and the end of software moats

00:15:00 Peak named user licensing and tokenization

00:17:40 Data sovereignty: why firms need to reshore their project data

00:22:20 Open SDKs and AI-written software

00:29:20 BIM 1.0 as a drawing conduit

00:38:30 Institutional knowledge drain and the people problem

00:41:10 The apprenticeship failure and AEC's next generation

00:50:40 Value-based billing and commoditization of design

00:54:40 Small teams competing with the largest firms

01:00:10 Autodesk's 2028 pricing wall

01:07:30 Mining your own data for competitive advantage

01:19:10 Token costs at scale and LLM selection

01:28:00 Multi-agent systems and governance structures

01:40:50 Sketch to 3D and the future of design ideation

02:10:00 The "anti-AI firm" and the future of practice size

👍 Like this episode if you're tracking what AI will actually do to how architecture firms work

💬 Drop a comment: which part of the AEC stack do you think is most at risk?

🔔 Subscribe for more conversations on architecture and technology

🌐 Explore more at [https://trxl.co](https://trxl.co/)

#AECtechnology #architecturebusiness #BIMfuture #AIinarchitecture #Autodesk #AECindustry