Equity What past investment booms can teach us about AI

Investment booms rarely unfold in a straight line. Railway expansion, electrification and China’s industrial rise all attracted doubts. Capital spending often ran ahead of visible demand, while productivity gains took time to emerge. Investors questioned whether the infrastructure being built would ever generate an adequate return.

Yet these cycles persisted for longer than many expected. They did not advance smoothly, nor did every participant benefit. Their significance became clearer only as adoption spread and the infrastructure found wider economic use. In time, each reshaped economies and investment markets in ways that had been difficult to anticipate at the outset.

Artificial intelligence may now be entering a transformative investment cycle of its own. Spending on data centres and computing capacity has accelerated sharply with power requirements rising alongside it. This has intensified the debate over whether future revenues can justify the amount of capital being deployed.

It is a valid concern. However, previous investment booms suggest that judging a structural cycle through near-term returns alone can provide an incomplete picture. The duration of a cycle matters too and should not be underestimated.

Adoption can begin gradually before gaining momentum as practical applications emerge. Financial returns may develop unevenly during this process. Market leadership can also change, particularly when new constraints alter where investment is needed and where value is created. Thus, periods of volatility do not solely signal that the underlying technology cycle has ended.

This historical perspective offers investors a useful framework for assessing AI. Rather than seeking a single verdict on valuations or near-term returns, investors can consider where the cycle may stand today. Five signposts are particularly relevant: the breadth of adoption, evidence of productivity and monetisation, the movement of bottlenecks, the supply response and the wider financial environment.

Combined, these indicators can help investors assess whether the AI cycle is still expanding or beginning to mature.

Viewed together, these signposts suggest that the AI opportunity could continue to evolve. As bottlenecks shift, investment may flow towards different parts of the supporting infrastructure. This could widen the opportunity beyond the companies that initially attracted investors’ attention and create potential beneficiaries across a broader range of sectors and geographies.

Historical comparisons should nevertheless be treated as a framework rather than a forecast. AI is developing more rapidly than earlier forms of infrastructure, and the long-term economics of models and applications remain uncertain. Markets can also respond quickly once a constraint is identified. Competition, regulation, capital intensity and financial conditions will all influence how the cycle progresses.

The central challenge is not predicting every market turn. It is determining whether the AI cycle is still expanding or beginning to mature. Examining the lessons of previous booms may help investors interpret volatility, identify where value could migrate and recognise the signals that matter most. 

To read more, download the full article. 

This insight is part of our broader analysis on how today’s global shifts are impacting investment opportunities – a dynamic we call The Great Global Restructuring.