Investors have become so focused on AI software that many are overlooking the industries making the boom possible.
I believe this could prove to be an expensive mistake.
The dominant investment narrative has revolved around the companies building AI models, designing advanced chips and providing cloud-computing services.
Investors have poured huge amounts of capital into firms such as NVIDIA, Microsoft, Amazon, Alphabet and Meta Platforms.
Those businesses have generated extraordinary returns and attracted vast amounts of capital. Yet the next phase of the AI story may have less to do with software and more to do with the physical infrastructure required to support it.
The scale of investment now taking place is staggering. Amazon, Microsoft, Alphabet and Meta are expected to spend more than $300 billion this year on capital expenditure, much of it directed towards AI infrastructure.
Across the industry, estimates for AI-related spending over the coming years run into the trillions.
Investors, understandably, focus on the headline beneficiaries. Nvidia has become one of the most valuable companies in the world. The hyperscalers dominate market indices. Every breakthrough in AI capabilities generates another wave of enthusiasm.
Less attention, however, is being paid to what all this investment actually requires.
Every new data centre demands enormous quantities of electricity. Every expansion of computing capacity requires transformers, switchgear, transmission equipment and cooling systems. Every upgrade to power infrastructure requires copper, aluminium and other industrial materials.
AI is rapidly becoming one of the largest infrastructure buildouts of the modern era.
This is where the investment opportunity starts to broaden.
The market has already recognised the value of the companies creating AI applications and supplying the most advanced semiconductors. Valuations in many parts of the sector reflect that recognition.
By contrast, many of the industries supporting the buildout remain underappreciated despite being critical to its success.
Copper is, for me, perhaps the clearest example.
Demand for the metal was already being driven by electrification, renewable energy projects and electric vehicles.
AI adds another powerful source of consumption. Data centres are highly copper-intensive, while the expansion of electricity networks needed to support them requires substantial additional supply.
Several industry forecasts suggest copper demand could outpace available supply later this decade. If those projections prove accurate, pricing power may increasingly shift towards producers at a time when new mining projects remain difficult, expensive and time-consuming to develop.
Power generation presents another area where investors may be underestimating future demand.
The International Energy Agency (IEA) expects electricity consumption from data centres to rise sharply over the remainder of the decade as AI adoption accelerates. Technology companies are already securing long-term energy agreements and exploring direct investments in generation capacity because reliable access to electricity is becoming a strategic priority.
Utilities were, for may years, often viewed as defensive investments, but AI may begin to change that perception.
Electricity demand growth has been relatively subdued across many developed economies for decades. The emergence of AI data centres as major consumers of power creates a different backdrop.
Utilities, independent power producers, grid operators and energy infrastructure businesses could find themselves benefiting from a structural demand trend that extends well beyond traditional economic cycles.
The same argument applies to industrial companies supplying the equipment needed to expand and modernise electricity networks.
Data centres cannot operate without power. Power cannot be delivered without transmission infrastructure. Transmission infrastructure cannot be built without specialised equipment that is already experiencing growing demand.
Investors should remember that major technological revolutions rarely create winners only among the companies developing the technology itself.
Railroads created opportunities in steel and industrial manufacturing. The growth of the internet generated significant value for network operators and infrastructure providers. The smartphone revolution benefited semiconductor manufacturers, component suppliers and communications companies.
AI is unlikely to be any different.
Markets often become captivated by the most visible beneficiaries while overlooking the businesses supplying the foundations upon which growth depends.
Today, AI is widely viewed as a software story. Increasingly, it also looks like a commodities, power and infrastructure narrative to me.
That distinction matters because some of the strongest investment opportunities over the coming decade may emerge outside the names that currently dominate headlines.
Investors don’t need to reduce exposure to AI leaders to benefit from this trend.
However, they should consider whether portfolios are overly concentrated in the obvious winners while ignoring sectors that may experience equally powerful demand growth.
As hyperscalers pour hundreds of billions into computing capacity and data centres, the next chapter of the AI investment story is, I believe, going to belong to the companies supplying the energy, materials and infrastructure that make it all possible.
Nigel Green is deVere CEO and Founder
Also published on Medium.
Follow Arabian Post
Select Arabian Post as your preferred source on Google and MSN News for trusted business news and Arab politics and updates.
