Taking stock of the AI boom

Artificial intelligence is no longer simply a technology story. It is increasingly an economic story, an infrastructure story and an investment story.

Today, AI sits at the centre of one of the most significant investment cycles in decades. From semiconductor manufacturers and cloud providers to data centre operators, utilities and industrial businesses, AI is reshaping corporate investment and economic activity. AI is not only helping people work more productively; for some cognitive tasks, it is beginning to provide capabilities that previously had to come from people.

The impact has been substantial. While a small number of large technology companies have driven a significant proportion of recent equity market gains, AI-related demand has also supported semiconductor manufacturers and technology exporters across Asia, particularly Taiwan and South Korea. At the same time, companies continue to accelerate spending on the computing power, networks and data centres required to support growing demand.

One of the biggest questions surrounding AI over the past year has been whether companies could generate sufficient returns from the enormous sums being invested in data centres, specialised chips and computing infrastructure. Investors questioned whether hundreds of billions of dollars of spending would ultimately translate into attractive revenues, profits and cash flows. Recent earnings announcements suggest those concerns are beginning to ease.

Recent results from the largest cloud providers provide some of the clearest evidence yet that AI monetisation is underway. Google reported record cloud revenue growth and substantial growth in contracted future orders, while Amazon Web Services delivered its fastest growth rate in several years alongside rising demand for both traditional cloud computing and AI-related services. Microsoft reported strong Azure growth, noting that newly deployed AI capacity was being monetised almost immediately as customer demand continued to exceed available supply. In each case, the evidence suggests that newly built infrastructure is increasingly generating commercial returns rather than sitting idle.

Businesses are increasingly embedding AI into core operations rather than simply experimenting with the technology. What began as a technology upgrade is becoming a broader business transformation story. Regardless of whether demand ultimately comes from OpenAI, Anthropic, enterprise software providers or open-source models, the common requirement remains computing power, storage and networking infrastructure.

Management teams are becoming increasingly confident about the potential returns these investments can generate.

Google, Amazon, Microsoft and Meta continue to increase spending plans, collectively committing hundreds of billions of dollars to data centres, specialised chips and AI infrastructure. Investors are increasingly weighing strong evidence of monetisation against the continuing need for substantial capital investment.

In many respects, AI is beginning to resemble an industrial investment cycle rather than a narrow technology story. Just as previous generations invested in railways, electricity networks and manufacturing capacity, today's technology leaders are investing in data centres, semiconductors, energy infrastructure and digital networks that may support future economic growth. Part of this spending may therefore be better understood as investment in new productive capacity rather than simply another generation of software. Strong structural themes do not automatically translate into attractive investment returns, particularly if valuations become overly optimistic.

Despite these encouraging developments, investors should recognise that risks remain. A meaningful share of AI-related demand is still concentrated among a relatively small number of customers. Industry estimates suggest that leading AI developers such as OpenAI and Anthropic account for a significant share of incremental demand for advanced computing infrastructure. While current demand remains strong, its long-term durability will depend on whether these businesses can continue attracting customers, generating revenues and moving towards profitability.

Recent results have also highlighted an important distinction between revenue growth and cash generation. Although revenues remain strong, free cash flow has come under pressure across much of the cloud-computing universe as investment spending continues to accelerate. The key question is therefore not simply whether AI demand remains strong, but whether future profits and cash flows will ultimately justify today's extraordinary levels of capital expenditure.

Importantly, durability of the AI build-out and attractiveness of individual investments are not necessarily the same thing. Underlying infrastructure expansion may prove highly durable even if some of the companies financing it ultimately deliver disappointing returns. History shows that periods of innovation are often accompanied by excessive optimism and speculative investment, reinforcing the importance of valuation discipline and diversification.

One of the most interesting developments is that the eventual winners may emerge from a much broader group of companies than many initially expected.

China provides a useful example of how the opportunity may broaden. Increasingly capable lower-cost and open-source AI models, as witnessed with DeepSeek suggest that businesses may benefit from AI without matching the investment made by the largest US platforms. While the United States continues to lead in developing many advanced systems, other regions may generate attractive returns by adopting AI rather than bearing the full cost of creating them.

Historically, technology offers a useful perspective. When the iPhone was launched, advanced processors represented a relatively small share of its value. Over time, software and computing capability became increasingly important. AI may follow a similar path, becoming progressively embedded within products and services and accounting for a growing share of economic value creation.

Perhaps more encouragingly, the benefits of AI are beginning to spread beyond companies building the technology itself. Businesses are increasingly reporting productivity improvements, cost savings and efficiency gains resulting from AI adoption. The conversation is gradually shifting from who is building AI to who can use it most effectively.

The employment implications are less clear. Some companies are adapting hiring plans and workforce structures, but economy-wide data does not yet show broad AI-driven job displacement.

There are already signs that this next phase is underway. AI is moving beyond software and cloud computing into areas such as power generation, electrical equipment, cooling systems, industrial infrastructure, logistics and real estate. In many respects, AI increasingly resembles an industrial investment theme rather than a narrow technology story.

For diversified investors, the key question is no longer whether AI matters but how best to gain exposure.

The first phase of the AI cycle was dominated by a relatively small number of large US technology companies. While these businesses remain central to the theme, relying solely on a narrow group of stocks introduces concentration risk. Future winners are also likely to emerge from other parts of the ecosystem, including semiconductor equipment providers, data-centre infrastructure specialists, utilities, industrial businesses, software developers and companies successfully using AI to improve productivity.

For portfolios such as PruFund, this argues for maintaining exposure to the AI theme without becoming overly reliant on a handful of US mega-cap technology companies. AI-related demand has also supported strong performance across parts of Asia, particularly Taiwan and South Korea through their exposure to semiconductor manufacturing and technology supply chains. Our approach seeks exposure across the wider AI ecosystem, including infrastructure providers, businesses enabling adoption, selected private-market opportunities and equity strategies designed to reduce reliance on market-cap concentration.

The evidence increasingly suggests that the technology is progressing from a phase dominated by investment and promise towards one characterised by adoption, monetisation and productivity gains. While the United States remains central to the story, future winners are likely to emerge across a broader ecosystem of semiconductor suppliers, infrastructure providers, industrial businesses, software developers and companies successfully applying AI to improve productivity.

The history of innovation suggests that transformative technologies create enormous economic value, but not every company benefits equally and investors can sometimes overpay for growth. The AI theme can succeed even if some AI-related investments disappoint.

For investors, the challenge is no longer simply finding exposure to AI. It is finding exposure in a diversified manner while remaining disciplined on valuation and concentration risk. Transformational technologies rarely create value in just one place, and the greatest opportunities are often captured across an ecosystem of beneficiaries rather than by a single company or market.

This content has been prepared by the Life Investment Office (LIO) for information purposes only and does not contain or constitute investment advice.