MAIN ARTICLE
A digital boom runs into physical constraints
Sabina Chiang, Standard Chartered’s head of managed investments and Hong Kong wealth solutions, described a central paradox of the AI build-out. Digital services still require decidedly physical inputs: uninterrupted electricity, high-capacity connections, cooling equipment, suitable sites and financing. As computational demand expands, infrastructure bottlenecks may offer opportunities to businesses able to relieve them. But scarcity alone does not guarantee good returns.
Chiang called this transition “Infrastructure 2.0,” a shift from a purely defensive allocation toward assets that may offer both resilience and growth. Data centers, telecommunications towers, fiber networks, batteries, nuclear generation, solar power and electricity grids are among the possible beneficiaries. Their financial characteristics, however, are quite different. An operational facility with contracted tenants is not equivalent to a speculative development awaiting permits, and an electricity project without long-term offtake agreements faces greater price risk.
Traditional discipline therefore remains essential. Is an asset necessary, difficult to reproduce and capable of producing relatively predictable cash flows? Can its contracts adjust for inflation? What is the cost of debt, and how concentrated are its customers? A project serving AI can still be a poor investment if its purchase price assumes flawless expansion, or if execution delays consume its expected return.














