For most of the twentieth century, the structural story of wood ended at about four storeys. Solid sawn timber is strong along the grain and weak across it; it splits, it burns, it varies tree to tree. Then someone glued the variance out of it. Cross-laminated timber (CLT) stacks lumber boards in alternating 90° layers — three, five, seven plies — and presses them into a panel that behaves like a two-way structural plate. The cross-lamination cancels the weak axis. What was a plank becomes a slab.

That single move is why an 18-storey building in Brumunddal, Norway, and a 25-storey tower in Milwaukee now stand on engineered wood. The signal worth reading in 2026 is not “wood is back.” It is that the load path changed shape. A CLT floor spans both directions, ties into glulam columns and beams through self-tapping screws and steel connectors, and resolves lateral wind through CLT shear walls or a concrete core. The mid-rise is no longer a frame wrapped in cladding — it is a system of pre-fabricated plates whose tolerances were decided in a factory, weeks before the crane arrived.

←TODAY: In 2026 a five-ply CLT floor panel routinely spans 5–7 m at residential load while storing roughly a tonne of CO₂ per cubic metre of wood. →3012: The Zurich-3012 city is built from materials whose supply graph a single architect can still draw by hand. Fulcrum: CLT only carries a mid-rise because the variance was moved off the building site and into a glue line — and that is exactly where the new single point of failure now lives.

The hidden topology: where the load — and the risk — actually flows

My desk reads buildings as dependency graphs, so look at where this one queues. The structural argument for CLT is clean: rolling shear in the cross-plies, not bending strength, is usually the governing limit state, which is why panels are sized by stiffness and vibration long before they run out of capacity. PAZ has covered the computational side of this thread before — the work on cost-effective timber shells through integrated computational design and mechanical half-lap joints in the archive shows how far the geometry-to-fabrication chain has matured. The joint is the project.

But the resilience engineer in me flags the supply topology. A CLT building depends on a forest, a glulam press, a CNC line, and a logistics window — in that order, with almost no slack. Most of Europe’s structural panel comes from a handful of large presses in Austria and the Nordics. That is a beautiful efficiency and a quiet single point of failure: one fire at one plant, one beetle outbreak in one supply forest, and a region’s mid-rise pipeline stalls. The genuinely interesting counter-move is decentralisation. UC Berkeley researchers, as ABC7 News reported this year, are pressing mass timber from wildfire-salvage wood — turning a fire-risk liability into local structural stock. That is not a sustainability footnote; it is a second node on a graph that currently has one.

Atelier: If your office takes on a mid-rise CLT project, draw the connection schedule before the floor plan — the screws, brackets, and panel edges decide the build sequence, and the architect who specifies them owns the tolerances. Treat the glulam supplier’s lead time as a structural input, not a procurement detail.

Hack: This Hack teaches you to sanity-check a CLT floor’s self-weight-to-capacity ratio in five lines before you trust a vendor span table. The domain is structural physics: a simply-supported panel’s mid-span moment under uniform load is wL²/8, which you compare against the panel’s bending resistance. Run this for your span and load and you will know in seconds whether you are stiffness-bound or strength-bound.

L = 6.0          # span (m)
w = 4.5e3        # design load (N/m, per metre width: dead + live)
M_ed = w * L**2 / 8           # design moment (N*m)
M_rd = 28e3                   # 5-ply panel resistance (N*m, from EN datasheet)
print(f"utilisation {M_ed / M_rd:.0%}  ->  {'OK' if M_ed < M_rd else 'oversize'}")

If utilisation lands near 40–60%, you are almost certainly governed by deflection and floor vibration, not bending — which tells you to reach for a deeper panel or a shorter span, not a stronger one.

El Takeaway de PAZ

CLT is not a greener cladding choice; it is a different load path with a different failure graph. The structural win — a two-way plate from a renewable, carbon-storing material — comes bundled with a concentrated supply chain that most spec sheets never show you. This week, do the Lin Rauch exercise on your next timber project: draw the real dependency graph from forest to press to crane, and find the third single point you didn’t know you had. Then design a second node into it.