The Gritty Reality of the AI Arms Race
The sheer scale of this build-out is difficult to comprehend. Training a single large language model requires a staggering amount of computational power, housed in data centres the size of aircraft hangars. This has kicked off what I believe is a capital expenditure supercycle. Companies are investing in tangible, physical hardware at a rate we haven’t seen in years.
This creates a ripple effect that extends far beyond the obvious players. The more data and AI models a company has, the more it needs to protect them, which is good news for cybersecurity firms. This entire ecosystem, from the chip designers to the server builders and network providers, represents the foundational layer of the AI economy. To me, investing in this theme isn't about betting on a single application, but on the entire digital bedrock. It’s a strategy that looks at the whole supply chain, much like the one found in the Oracle's AI Cloud Catalyst basket, which groups these essential infrastructure players together.
Of course, no investment is without its potential pitfalls. This is a fiercely competitive space, and today’s technological marvel could be tomorrow’s museum piece. Valuations for many of these stocks are, shall we say, optimistic, and they could be punished severely if growth slows. But the fundamental trend seems undeniable. AI is not a fad, it’s a foundational technology shift. The companies building the infrastructure to support it are likely to be busy for a very long time to come.
