The "picks and shovels" AI play offering true diversification

This interview was filmed on Monday 2 March 2026.

While the debate over its long-term impact rattles on, AI has undeniably taken on the mantle of kingmaker when it comes to equities markets.

Entire sectors are already rising and falling simply on the perceived impacts of AI, but for investors looking for a less volatile way to play this latest all-consuming megatrend, the answer could be fairly simple.

As part of Livewire's 2026 Listed Series, I spoke to Global X's Billy Leung about the new leg in the AI cycle and the unparalleled opportunity it's offering investors.

Phase Three of the AI supercycle

While AI has certainly dominated market discussions since arguably the launch of ChatGPT in November 2022, the technological evolution stretches back further.

Six or seven years ago was "phase one" of the AI cycle, where compute was the central driver. Then came "phase two", in which the hyperscalers - including many of the Mag 7 and private companies like OpenAI and Anthropic - entered the arms race. This is now giving way to a new phase, says Leung.

"Right now the current phase that we're seeing is the AI infrastructure phase (or "phase three"), which is probably where we think most of the investment opportunity is going to be in the next few years."

In phase two, the investment opportunity was mostly limited to the hyperscalers themselves. Now in phase three, the investable opportunities are broadening out significantly.

"In terms of this AI infrastructure phase, we see it as more multifaceted and multi-layered," says Leung. "We don't see this as simply just data centres. We see it as the picks and shovels that go into this infrastructure build out."

"That includes building the cables that go in the data centres, the cooling equipment, the energy storage system equipment that goes in the data centre, to even the raw materials and the energy that is needed to power the data centres."

Where the investment opportunity is

As history attests, the simplest way to play any transformational technology remains the "picks and shovels" approach. And that's the philosophy Global X has taken on board.

The Global X AI Infrastructure ETF (ASX: AINF) offers exposure to 30 global companies across the wider infrastructure thematic.

"It encompasses the picks and shovels that goes into the infrastructure," says Leung.

There are 10 data centre infrastructure names, 10 covering the energy and cooling equipment requirements and 10 related to the raw materials needed, mostly copper and uranium.

Unlike phase two, where a bet on AI was effectively a single concentrated bet on US Big Tech, phase three of the AI cycle means genuine diversification across regions, sectors and stocks.

"I think with this AI infrastructure theme, it's really important to understand that it goes further than just one single type of company."

Instead of trying to pick the winners in the AI arms race, the AI infrastructure phase is more akin to an agnostic bet on the technology itself, and that lends itself to the ETF structure, says Leung.

"It is the entire value chain that we have to really consider and why I think our AINF ETF really appeals to investors, because it allows them to capture the entire data centre AI infrastructure build out within just one product."

Leung identifies a few stocks within the ETF that encapsulate the opportunity for investors and the diversification it offers.

The first is Vertiv (NYSE: VRT), an energy equipment company that is now one of the major providers of cooling equipment for data centres globally.

Another is Siemens AG (ETR: SIE), which has transitioned from telecommunications to automation and now helps power much of the data centre sector.

There's also Southern Copper (NYSE: SCCO), one of the world's largest copper miners.

Where to tread carefully

Global X is optimistic on AI as a legitimately transformational technology with much more to run. "We believe that it's a long-term structural story," says Leung.

"If you look from a perspective of the current investment as a percentage of GDP and the potential adjustable market as a percentage of total global GDP, AI is still by far one of the lowest in terms of actual investments compared to previous historical innovations."

But he does urge some caution around excessive expenditure. In what's likely to be near-enough a winner-takes-all market, Leung says investors need to be wary of companies, including Mag 7 hyperscalers, that may be throwing too much money at AI in anticipation of securing market share that may not materialise. "

AI is not over-invested, but there might be some specific companies that are miscalculating their future market share, and which could lead them to over-invest."

It's created a crowded trade in US big tech, while better opportunities in Asia and humanoid robotics are being overlooked.

But if AI is a story that's been hard for investors to ignore over the last few years, there's good reason for that, says Leung.

"One of my colleagues in the US always said that there's only going to be two kinds of companies in the future. Ones that use AI and ones that don't exist, and we're strong believers of that."