27 Feb 2026: Platform Reorgs, Starts Illusions, Skills Reality, Precon AI, and Data Centre Politics

Five signals sitting underneath delivery. Autodesk puts a real deadline on a platform regroup. US starts look stable until you remove three giant energy jobs. The UK skills gap is structural, not cyclical. A preconstruction AI startup wants planning to be intense so site work can be boring. And data centres are discovering that local politics can shape pipelines as much as capital.

01 · Platform & Workflow

Autodesk folds Construction Cloud into Forma

Autodesk says Construction Cloud formally joins Autodesk Forma on 24 March 2026. Product names change. The CDE layer becomes Forma Data Management. Autodesk says projects and integrations stay intact. This is more than a rebrand. It shifts how the platform presents itself from design through construction and operations.

Hook: If you sell into the Autodesk ecosystem, every doc, deck, API reference and onboarding flow now has a deadline. Are you ready for 24 March?

02 · Built Environment Economics

The US starts headline hides concentration risk

US total starts ticked up in January. But three energy megaprojects drove almost $20B of that growth. Strip them out and residential, nonresidential, highway and bridge work all fall. Electric power and utilities surged. Highways dropped 42.3 percent month over month.

Hook: If you are not tied to LNG, power or hyperscale, the average number is not your reality. How exposed is your pipeline to one giant job slipping?

03 · Workforce

The UK skills gap is structural

CITB says the UK needs more than 47,000 additional construction workers each year through 2029. Construction already has the highest density of hard to fill vacancies in the country. This is not a cycle problem. It is a pipeline problem. When demand ramps, the same trades get pulled into the same programmes.

Hook: When spending accelerates, who actually has the people to deliver it?

04 · Preconstruction AI

Make planning intense so delivery can be boring

MeltPlan is building an AI planning engine focused on code, cost and schedule trade offs before site work begins. They are piloting with DPR in California and Innovo in the UAE. The argument is simple. Most failures start upstream in fragmented preconstruction decisions.

Hook: Does it reduce decision latency for estimators and PMs, or just create another review layer?

05 · Regulation

Data centre growth meets local resistance

Communities are pushing back on data centres over water use, energy demand and limited local jobs. New York is proposing a three year moratorium on new permits. New Orleans has paused permits for a year. The constraint is no longer just capital or EPC capacity. It is social licence and permitting.

Hook: How many sure thing hyperscale projects quietly become political risk?

The thread

Delivery is getting squeezed from both sides. Platforms are consolidating workflows. Starts are concentrated in a handful of giant jobs. Labour remains tight even in choppy markets. AI tools must prove they fit real precon workflows. And the hottest sector in construction now faces local political friction. The common theme is not hype. It is constraint.

One practical move this week

Run a 30 minute delivery risk audit on one live project. What platform or data handoffs could break in 30 days. Where is your labour plan brittle. Which permits could slip. Assign a name to each risk. If it matters, it needs an owner.

Want the full picture

Every source. Deeper context. The bits being politely ignored.