Rapid expansion of artificial intelligence infrastructure is creating growing uncertainty around how construction-related expenses are recorded in company financials, adding complexity for investors assessing long-term risks. Large technology firms are committing massive capital to build AI-ready data centres, but the way these costs are reported often lacks clarity.
Companies such as Alphabet, Amazon, Meta Platforms and Microsoft are investing heavily in specialised facilities to support AI workloads. Much of this spending is grouped under broad accounting categories like construction-in-progress, which combine diverse elements into a single line item.
The challenge lies in the nature of AI infrastructure itself. Long-lasting assets such as buildings and power systems are often recorded alongside high-cost computing equipment and chips that can become outdated within a few years. Analysts say this approach makes it difficult to judge how much capital is tied to durable infrastructure versus fast-depreciating technology.
While companies regularly highlight the scale of their AI ambitions, detailed breakdowns of construction and equipment costs are rarely disclosed. Accounting rules currently do not require firms to separate these components in detail, leaving investors with limited visibility. As AI-driven construction accelerates, calls for more transparent reporting are expected to grow louder.
