Nvidia’s return to the top of the market has amplified confidence in the generative AI theme, with analysts now pointing to a structural shift in how investors are viewing the sector.
A new phase of AI monetisation appears to be emerging, with some forecasters predicting AI data centre spend could reach US$2 trillion by 2028.
Billy Leung, technology analyst at Global X, said the firm remains “constructive” on AI, citing signs that the market narrative is evolving.
“We remain constructive on AI, particularly now that the narrative has started to evolve from AI buildout to AI monetisation,” Leung told InvestorDaily.
“The past few weeks have been telling. Microsoft’s Azure growth reacceleration, Snowflake’s customer enthusiasm around Cortex, and Nvidia’s sovereign AI wins in Europe all signal that the theme is entering a more mature and revenue-accretive phase.”
Leung noted that market participation has broadened significantly compared to last year.
“Last year, investors clustered into a narrow set of names. This year, software is catching up, optical hardware is reawakening, and sovereign demand is emerging as a new driver alongside hyperscaler capex,” he said.
“It’s not just about GPUs anymore as we’re also watching infrastructure, tooling, and orchestration layers get repriced in real time.”
On suggestions that the Magnificent Seven have peaked, Leung pushed back.
“The idea that their run is over doesn’t quite hold up. Some consolidation is natural after such a strong stretch, but the underlying fundamentals (like Microsoft's AI-linked cloud growth or Nvidia's margin upside) remain intact,” he said.
“If anything, we’re seeing leadership rotate within the group rather than fade. Apple may be lagging for now, but others are stepping up and the further earnings rounds could extend that divergence.”
Similarly, Betashares’ investment strategist, Hugh Lam, said the Mag 7 have consistently defied sceptics and delivered strong earnings.
“In our view, Betashares remains constructive on the Mag 7 given their importance in providing critical computing infrastructure that allow enterprises to both train and infer AI models. Earnings have grown 16 per cent in the first quarter of this year and six of the seven companies have reported positive EPS surprise,” he explained.
Looking ahead, both Leung and Lam see significant long-term potential for the AI theme, especially if sovereign AI adoption accelerates.
“If sovereign AI adoption follows the same trajectory as hyperscaler spend, and if tools like Snowflake’s Cortex and Microsoft's Copilot continue to gain enterprise traction, then we’re only just scratching the surface,” Leung said.
“For long-term investors, having exposure across the full stack semis, infrastructure, and software is increasingly important.”
Lam added that the collective goal is superintelligence, which should continue to support capital investment across the AI ecosystem including in areas such as computing infrastructure, cyber security, defence, and robotics.
“As the threat landscape evolves, security vendors like CrowdStrike and Palo Alto are incorporating AI within their offerings to automate the threat detection process and defence companies like Palantir have been instrumental in providing AI-powered military systems for the US government to advance their military agenda,” he said, adding that physical AI is another growing theme.
“Particularly in robotics which is expected to be a US$5 trillion market by 2050 according to research by Morgan Stanley.
“Over time, AI technologies should help robot software models become more sophisticated and cheaper, and when integrated tightly with hardware presents a huge market for potential growth.”
