Artificial intelligence seems to be the only game in town these days, and to a certain extent, it is. I believe AI represents one of the greatest technological advancements in history and an investment opportunity that will unfold over decades. But what exactly is an AI company, and how should investors approach the opportunity?
There are the chipmakers developing the processors that power AI, the hyperscalers such as Microsoft, Amazon and Google, the companies developing AI models and the robotics businesses that will eventually bring AI into the physical world. Then there are the energy providers supplying the electricity needed to run it all, and the businesses that will benefit from adopting AI to improve their own operations. The opportunity is enormous, but not all these businesses are created equal. Some offer compelling investment opportunities, while others carry substantial risks. AI is too broad and diverse to treat as a single investment category. Investors need to understand the different layers of the industry and the quality of the businesses operating within them.
One group attracting attention is the emerging neocloud sector. Companies such as CoreWeave, Firmus and Nscale provide specialised computing infrastructure that helps power AI. Their emergence reflects an imbalance between demand for computing capacity and the industry's ability to supply it. Even the world's largest technology companies are struggling to build infrastructure quickly enough to satisfy that demand. Increasingly, they are turning to specialist providers to bridge the gap.
But being essential to the AI revolution doesn't necessarily make them good investments.
The recent withdrawal of Australian company Firmus's proposed IPO highlights the risks. Its proposed valuation raised serious questions, particularly given the potential for existing shareholders to sell down their holdings. In South Korea, neocloud provider Elice recently listed following high retail demand, only for its shares to plunge 32% on debut. The stock has since recovered somewhat but remains below its issue price. Meanwhile, British provider Nscale recently had Microsoft withdraw as a customer for one of its major projects after delays in securing electricity supply. These are businesses making huge commitments to infrastructure that can take years to deliver.
Another concerning example is CoreWeave. Its massive growth has been fueled by substantial debt, and the bond market is beginning to question the risks. The yield on its debt maturing in 2032 recently approached 15%, reflecting growing concerns about its financial position. For a business dependent on continued access to capital, that really matters. Higher borrowing costs can quickly undermine the economics of new projects, while an inability to refinance existing debt could threaten the survival of the more heavily leveraged operators.
There is an interesting dynamic at play here. The hyperscalers are investing enormous amounts in their own infrastructure, but increasingly they are relying on other companies to finance and build additional capacity. In doing so, they are transferring some of the capital requirements and financial risks to businesses with considerably weaker balance sheets. The hyperscalers still have substantial financial commitments, but they also have highly profitable established businesses supporting their investment. Many neoclouds do not.
For our clients, the distinction is important. We want exposure to the extraordinary long-term opportunity in AI without taking unnecessary financial risks. That is why we continue to favour established businesses such as Microsoft, Amazon, Alphabet, and Meta, alongside energy producers and, increasingly, businesses that will benefit from adopting AI.
AI will transform the global economy over the coming decades. But not every company helping to build that future will be a good investment.
General Advice Disclaimer: This information is of a general nature only and may not be relevant to your particular circumstances. The circumstances of each investor are different, and you should seek advice from an investment adviser who can consider if the strategies and products are right for you. Historical performance is often not a reliable indicator of future performance. You should not rely solely on historical performance to make investment decisions.