After a year dominated by geopolitics, 2026 is shaping up as the year the AI trade is tested by markets, not hype.
At the World Economic Forum in Davos, global leaders focused less on inflation or elections and more on whether AI is delivering real economic value, or inflating the next great market bubble. Confidence remains high. Nvidia CEO Jensen Huang described AI as a “five-layer cake” capable of triggering the largest infrastructure build-out in history, while Anthropic’s Dario Amodei spoke of AI unlocking unprecedented productivity.
Markets have embraced the narrative. AI-linked stocks have dominated returns, capital spending runs into the trillions, and Nvidia’s GPUs have become critical choke points in global supply chains.
Yet doubts are emerging. Microsoft CEO Satya Nadella warned that if AI’s benefits fail to spread beyond tech, it would signal a bubble. JPMorgan chief Jamie Dimon has cautioned that AI’s rollout may be moving faster than society can absorb.
That tension is increasingly visible in markets. Volatility is rising, with valuations pricing in near-perfect execution and future profits that have yet to materialise. This matters globally: just ten US tech stocks account for around 40% of the S&P 500, leaving investors, including Australian super funds, highly exposed.
Australia now sits at an unexpected crossroads. The expansion of data centres, driven by demand for “AI tokens”, units of computing power, could position the country as an exporter of AI infrastructure. But the trade-offs are real: power-hungry facilities, rising energy costs and job disruption.
For investors, the AI story isn’t over, but it’s becoming riskier. The key question for 2026 is simple: how much of the AI future is already priced in?
