Tech Hits Turbulence

AI infrastructure is entering a more disciplined build-out that should pressure speculative projects while preserving the long-term value of approved, power-secured data centers.

• The November 6 commentary links the market sell-off to the prospect of greater scrutiny of AI and data-centre development after the U.S. midterm elections, alongside existing concerns about technology valuations and hyperscaler capital spending.
• Aggressive forecasts had placed AI-related capital expenditure at $1.5 trillion to $2 trillion in 2027, while spending closer to $1 trillion would ease supply-chain pressure and encourage more selective allocation.
• Permitting, power, chips and cooling systems are limiting how quickly new capacity can come online, while local communities are seeking greater benefits from data-centre projects.
• Technology companies have increasingly relied on bond, credit and equity issuance to fund capital expenditure, so a slower investment cycle could reduce leverage, dilution and hyperscaler debt supply.
• Frontier-AI campuses requiring hundreds of megawatts face greater infrastructure and political constraints than low-latency cloud facilities, whose demand remains supported by continued digitalisation.

Takeaway: Investors should distinguish between projects by their approvals, power access, community support, financing and end use, since regulatory friction may increase the scarcity value of existing digital infrastructure.