Where are the hottest opportunities to make money from AI?

Please note this interview was filmed on 13 March 2025.

Artificial Intelligence (AI) has had an explosive run since the release of ChatGPT in November 2022. For example, NVIDIA saw its stock soar from ~$14 to $140 (before the recent correction), a 10-bagger in just two years!

With AI-linked stocks taking a breather, investors are reassessing the best way to play the theme. The conversation is shifting towards AI’s real-world applications and monetisation. At the same time, concerns are emerging around the costs for AI companies, particularly energy demands and data centre infrastructure.

As a result, investors are now focusing on the second-order beneficiaries - the companies that stand to gain from AI adoption, rather than those creating the technology.

Billy Leung, Senior Investment Strategist at Global X, helps us navigate where to look for opportunities in the evolving AI sector.

In this episode of The Pitch, Leung discusses the sectors poised to benefit most from AI’s next phase, the risks to watch out for, and how Australian investors can position themselves to capitalise on future winners.

Edited transcript

AI had an explosive run in 2024, but the conversation is shifting towards monetisation and real-world application. Where do you see the next big breakthroughs happening?

One analogy I often share with clients is that traditional investing is linear . You buy a coal mine, extract the commodity, sell it, and that’s the value chain. Innovation, on the other hand, is circular. Hardware enables software, which creates infrastructure, which enables adoption and that cycle feeds itself.

Most of the AI focus so far has been on hardware. We saw strong growth in semiconductors and the equipment behind them. That made sense in the early stages. But now, with cheaper training and infrastructure coming through, we could see monetisation and adoption accelerate.

It reminds me of the early smartphone days. Adoption was slow until the "killer apps" - Instagram, Facebook, Angry Birds - sparked mass uptake. I think we’re nearing that moment for AI. I don’t know exactly what those "killer apps" will be, but they’re what will drive the next wave of adoption.

Data centres and energy infrastructure are becoming crucial for AI’s expansion. How should investors think about the interplay between AI and energy demand?

Even with more efficient training methods, energy demand remains a big concern. That’s why we’re seeing growing interest in alternative and renewable energy sources.

One area that’s become especially relevant is nuclear - specifically, uranium. Big tech names like Google, Amazon, and Meta are investing in or purchasing nuclear-powered infrastructure. Not only does nuclear produce low emissions, but advances in technology have made it safer, faster, and cheaper to deploy.

For context, traditional nuclear plants took 7–10 years to build. Now, small modular reactors (SMRs) can be built in 3–5 years, at a lower cost, and are often located underground for safety. This evolution in energy is making AI infrastructure more scalable and sustainable.

Let’s talk about AI adoption beyond technology. What sectors could be the biggest beneficiaries?

That’s the exciting part - figuring out what the next "killer apps" might be. 

One area I don’t think is getting enough attention is robotics. Historically, the challenge has been cost - it’s expensive to integrate AI with robotics. But with cheaper training and more efficient models, we could see robotics adoption accelerate quickly. Some major companies are already moving in this space.

Another underappreciated sector is cybersecurity. It’s long been seen as its own niche, but people don’t realise how reliant it is on speed, capacity, and processing power. 

AI significantly enhances all of that, allowing for faster, more scalable, and more cost-effective threat detection. As AI adoption grows, I expect cybersecurity to be a major beneficiary.

We’ve seen DeepSeek cause volatility by potentially lowering the cost of large language models. What are your thoughts?

We’ve always talked about the circular nature of innovation. Hardware drives infrastructure, which drives adoption. DeepSeek has fast-tracked that cycle.

I see two key implications: First, how companies spend will shift. We’ll likely see capital move away from high-cost training and towards inference and real-world deployment. That means more focus on delivering outcomes, not just building models.

Second, the beneficiaries could change. As adoption becomes the focus, companies with strong application layers and go-to-market capabilities stand to benefit. Interestingly, a lot of Chinese companies are excelling in this area - they’re strong developers and quick adopters. That’s why we’re already seeing some of them benefit from DeepSeek’s advancements.

Lastly, Billy, how can investors gain exposure to these AI trends right now?

So we've gone through and talked about how AI is very broad and there are a lot of opportunities and this is why we built the Global X Artificial Intelligence ETF (ASX: GXAI). 

GXAI covers the entire value chain of the AI space and includes the top 85 companies by market cap which are using AI, developing AI, or they're using the hardware that develops AI. So that actually has that whole value chain that we keep talking about.

The beauty of it is that we cap exposures at 3% to ensure we have a very good blend of all these companies across the AI spectrum including pioneers and adopters.

And that's been resonating a lot with clients, mainly because they understand they want to be exposed to AI, but it's very hard to pick which areas are the correct segments to be invested in any one time. 

For investors who want exposure but aren’t sure which segment to bet on, GXAI offers a diversified and intuitive way to participate in the AI megatrend.

ETF

Global X Artificial Intelligence ETF (GXAI)

Global Shares

View