Executive Briefing: Construction Is the Problem the AI Revolution Didn't See Coming

The world's largest asset manager just wrote a $100M cheque to train electricians. Google put up $50M. Microsoft named electrical talent as the single biggest challenge to building data centres. The firms selling the AI dream just discovered they can't deliver it without construction workers.

This week on the Executive Briefing:

• Why data centre companies are paying $150K starting salaries for skilled trades, and what that means for every contractor drawing from the same labour pool. The real numbers on construction robotics — $1.36B in venture funding, 125% up year on year — and the four workflows where robots are actually working, not just demoing.

• The Iran oil shock is a second front, not the first. US construction input prices were already running at 12.6% annualised before a single missile was launched. Aaron Anderson's analysis shows 130,000 fewer permits being filed. We break down which sectors are pulling back and which are ploughing ahead regardless.

• The AI question nobody is asking: the difference between deterministic and probabilistic, why every AI tool you're being sold is fundamentally a guess, and how to know which of your workflows can tolerate that and which ones can't. Plus Anna Berger from Trayd on why 99.7% accuracy still isn't good enough for construction payroll.

Quick hits: NEOM's Line is effectively dead, the UK government is abolishing retention payments, and a US jury just ruled social media platforms were deliberately engineered to be addictive.

Join the debate: