Governance, power and delivery are the blockers

A Revizto-backed survey reported by ITBrief says Australian architecture, engineering and construction firms now see regulation, privacy and accountability as bigger blockers to AI adoption than finding use cases. That is a useful signal. The conversation has moved on from "what can AI do?" to "what can we safely put into production?".

For construction teams, this is not an abstract policy point. If the permissions, data handling and sign-off rules are fuzzy, AI just speeds up the same confusion. The firms that get further will be the ones that sort the boring governance first.

The Axios piece on the Trump-branded Texas project is a reminder that the gap between AI ambition and construction reality is real. First buildings are slipping, capacity targets are moving and the CEO has gone. At the same time, another analysis argues that half of 2026's AI data centre projects will not get built because transformers, switchgear and grid connections are the real bottlenecks.

This is where the market gets useful for construction people. The demand is there, but delivery now depends on power, procurement and programme discipline. That is not hype. That is a built environment problem.

Applied Digital is still expanding its AI infrastructure footprint. The company reported strong revenue growth, started construction on Delta Forge 1 and leaned on the same capital-heavy model that is now common across the sector. The numbers are a good reminder that this is still a live build market, not a passing software trend.

That does not make it easy. It makes it capital intensive. But, if you work in construction, that means more demand for power, civil, M&E, commissioning and delivery control.