AI is a major theme among investors this year, according to Betashares investment strategist Hugh Lam, who cited substantial inflows into the exchange-traded fund (ETF) provider’s tech-focused thematic ETFs.
Yet, he warned, the sector remains highly volatile, as shown by Nvidia’s loss of over $500 billion in market capitalisation following the January release of the DeepSeek carbon model.
But the differences in returns among mega-cap tech stocks mean investors no longer see the Magnificent Seven as a uniform category, creating openings for emerging companies such as Palantir, which he described as the new “poster child” of the AI and defence sector.
Moreover, Lam highlighted that China’s equities are emerging as a compelling asset class following Xi Jinping’s pivot towards supporting the tech sector, including major players such as Alibaba and Tencent.
“It’s really come down to what’s happening at the policy level. With Xi, you know, he’s really embraced the vibrant technology sector through Alibaba, Tencent, all these major players,” he said.
“And if you think back a few years ago, throughout 2020 to 2022, that’s a remarkable change, when there was regulatory crackdown.”
The Hang Seng tech index is up 24 per cent compared with broad indices such as the S&P 500 and Nasdaq, Lam added, with growth accelerated by companies launching their own AI models.
Infrastructure and physical AI present investment opportunities
Beyond individual equities, panellists stressed opportunities in the AI infrastructure ecosystem.
Mike Zimmerman, partner at deep tech venture capital firm Main Sequence, said: “The first is the infrastructure to make AI run more efficiently. I think, probably, people know just how much data centres cost to power and how much computation and buildout is required to actually meet the demands of AI, so there’s a huge efficiency opportunity.”
Zimmerman noted that Australian researchers are also developing more energy-efficient architecture, while the country maintains a strong track record in software-as-a-service, with companies such as Dominic Global, Atlassian and Canva leading the way.
He sees this culminating in specialised applications, including AI-powered insurance solutions, and emphasised Australia’s potential in physical AI, drawing on heavy industries such as mining and construction for training data and robotics development.
“We actually have a number of world leaders in that space, companies like Emerson, that’s using computer vision to do autonomous mind mapping,” he said.
The investment landscape is further shaped by global policy and geopolitics.
Australia at a critical point
Associate professor at UTS, Dr Marina Yue Zhang, an expert in Australia–China relations, said Australia stands at a critical point in developing its AI strategy.
“Australia needs to find a way – and that is not just unique to Australia, for all middle power countries – to move forward in maybe the most important geopolitical battlefield for the times we can see in the next decades or even longer,” Zhang said.
Zhang advocated for the government to “play a more active role” in fostering sovereign AI capabilities, enabling collaboration between researchers and start-ups while sharing computing power and infrastructure.
Drawing lessons from the distinct AI strategies of global leaders China and the US, Zhang highlighted some of the advantages of China’s centralised government approach.
While in the US, large tech giants have dominated the sector, with substantial capital expenditure in data centres a significant driver of gross domestic product growth, Zhang said China is using a more “counter-intuitive” method.
“China actually is more of a bottom-up innovation in the sense that it is government-led, especially provincial government-led competition for building enormous numbers of data centres,” she said.
According to Zhang, China’s government-backed strategy enables projects to launch efficiently and compete on a global scale. Additionally, this rapid advancement is fuelled by free access to data centres and data, provided by the government, as well as the ability to scale applications and physical AI robotics.
On the other hand, Zhang said a key challenge for the US is the reluctance of its private tech giants to share their computing power with start-ups.
“So that kind of looks like the growth of tech giants is actually killing entrepreneurial spirit in the United States. So if everything is done by large firms, that is actually not good,” she said.
Zhang argued that Australia could benefit from emulating the Chinese model by not only removing regulatory obstacles – as has been a major discussion point at last week’s productivity roundtable – but by actively promoting government involvement.
“Australia is very strong in standards and Australia can continue to play a pivotal role in that area, if this government supports and has investments,” Zhang said, adding that the country has an important role to play in bridging the countries in the Asia-Pacific region with the mainstream AI players.
Ultimately, panellists said Australia faces both challenges and opportunities in the evolving AI landscape. While the country’s AI ecosystem is still developing, its strong software industry, capacity for specialised applications and potential to leverage physical AI in heavy industries provide fertile ground for investment.
“You know, we’re in August. We’ve got a few more months till the end of 2025 and yes, definitely, increasingly being dominated by AI,” Lam said.
