Last week’s Nvidia results were extraordinary even by AI standards. The company reported quarterly revenue of US$96.2 billion, up 106 per cent from a year earlier, while its data-centre business grew 117 per cent. Perhaps even more strikingly, Nvidia expects revenue to grow approximately 70 per cent in fiscal 2028. For a company already operating at Nvidia’s scale, those numbers underscore just how powerful the artificial intelligence investment cycle remains.
But perhaps the more important message for investors goes beyond Nvidia itself. The AI trade is starting to change. As spending accelerates, the investment opportunity is broadening beyond semiconductors and mega-cap technology companies to the infrastructure required to support the AI buildout.
The spending tells the story
The scale of the investment is difficult to ignore. Bloomberg consensus estimates project that major hyperscalers could spend roughly US$650 billion in 2026, US$750 billion in 2027, US$800 billion in 2028 and US$760 billion in 2029. That is close to US$3 trillion of capital spending over four years, with estimates continuing to move higher as companies expand their AI ambitions.
And that money has to go somewhere. You cannot build AI in the cloud without building an enormous amount of infrastructure on the ground. Every new data centre requires advanced semiconductors, but it also requires buildings, electricity, transmission capacity, transformers, cooling systems, networking equipment and backup power. In many respects, these are the picks and shovels of the AI buildout.
Perhaps most importantly, AI may live in the cloud, but the cloud runs on electricity. According to the International Energy Agency, global data-centre electricity consumption is projected to roughly double from 485 terawatt-hours in 2025 to approximately 950 terawatt-hours by 2030, while electricity consumption from AI-focused data centres is expected to grow even faster. In the United States, data centres are expected to become one of the largest sources of incremental electricity demand through the end of the decade.
For investors, that significantly broadens the opportunity. AI spending may increasingly benefit utilities, power producers, electrical equipment manufacturers, engineering and construction companies, cooling providers and other industrial businesses that would not traditionally be considered part of the technology sector. This is why I believe the AI trade is becoming both bigger and broader.
A powerful theme still requires discipline
Of course, enormous capital spending does not automatically translate into enormous investment returns. One of the increasingly important questions is whether the revenues, productivity gains and profits ultimately generated from AI will justify the amount of capital being committed today.
History offers plenty of examples of transformative technologies that changed the world while investors still paid too much for individual companies along the way. Valuation matters. Earnings growth matters. Free cash flow and return on invested capital matter. Competitive advantages can also change quickly in an industry evolving as rapidly as AI.
Diversification therefore remains particularly important. Investors can participate in a powerful secular growth theme without allowing one company, sector or investment narrative to dominate a portfolio. Owning five different AI-related stocks does not necessarily mean you are diversified if all five ultimately depend on the same underlying AI spending cycle.
The opportunity, in my view, is less about abandoning the established AI leaders and more about recognizing that the investment universe is expanding. The first phase was dominated by chips and mega-cap technology companies. The next phase could increasingly extend from chips to data centres, from data centres to power, and from power to the equipment and infrastructure required to support it all.
Keeping the bigger picture in mind
Even a compelling long-term investment theme needs to be considered as part of a broader portfolio. The appropriate level of AI exposure will vary depending on an investor’s existing concentration, risk tolerance, time horizon, liquidity requirements and longer-term financial objectives.
That is why investment decisions around AI, like any other major theme, should ultimately be made within the context of each investor’s broader total wealth plan. The goal is not simply to identify the next great investment trend, but to determine how that opportunity fits into a diversified portfolio designed to achieve longer-term financial objectives.
AI may prove to be one of the defining investment trends of this generation. The next chapter, however, may be less about finding the next Nvidia and more about identifying the companies supplying the power, equipment and infrastructure that make the AI revolution possible.
The AI trade is getting bigger. The opportunity may increasingly lie in looking broader.
