Investors became increasingly cautious around AI-related valuations as capital expenditure requirements and financing risks continued escalating.
Market sentiment turned more cautious last week, even as Nvidia, the bellwether of the technology sector, reported robust results. Its earnings release delivered a brief lift in optimism but failed to sustain equity-market momentum. Valuations remain stretched across much of the market, and investors appear increasingly wary. Questions are also mounting over the sheer scale of investment now being funnelled into artificial-intelligence infrastructure. OpenAI, whose annual revenues are in the region of US$15 billion - whilst operating losses continue to accumulate - has outlined capital investment plans of around US$1.5 trillion over the coming years. The magnitude is almost surreal. To offer perspective: at the rate of one dollar per second, US$15 billion equates to approximately 47½ years - roughly a working life - whereas US$1.5 trillion would take some 47,500 years, comparable to the period since modern humans first settled Europe! Clearly, much must go right for some companies at the centre of the theme.
Perhaps more important than the raw sums is how such investment is financed. In conversations last week with Xuesong Zhao, manager of the Polar Capital Artificial Intelligence Fund - held in the T. Bailey funds of funds since 2017 - the focus turned to the risks of debt-funded capital expenditure. The major hyperscale firms have typically financed investment from their impressive cash flows whilst smaller competitors have leaned more heavily on borrowing. Should AI supply expand faster than demand, firms with higher leverage are likely to feel the strain first.
A further challenge is how AI is monetised. Usage is accelerating rapidly, but most revenue models still rely on flat fees that bear little relation to actual consumption. The next stage in commercial development may come from Agentic AI - systems capable of autonomously performing complex tasks and integrating directly with business processes - which could pave the way for more lucrative, usage-based pricing. The early signs are encouraging in fields such as dynamic pricing, where firms such as Delta Airlines are using AI to set fares in real time, boosting both revenues and efficiency. As the investment theme shifts from those building the technology to those exploiting it, the Polar Capital Artificial Intelligence Fund continues to adjust its positioning and to guide investors who remain sceptical of the sector’s more sensational claims.
