AI-bubble chatter is getting hard to ignore. Even industry heavyweights – Sam Altman, Jeff Bezos, Mark Zuckerberg – are sounding cautious. So where does that leave ordinary investors trying to make sense of today’s market? For me, the answer lies in looking beyond the obvious.
AI expenditure
It’s impossible to pin down the exact amount Microsoft, Meta, and the other hyperscalers are spending on AI. But we do know it already runs into the hundreds of billions – and it’s still accelerating at a record pace.
History tells us this pattern isn’t unusual. Whenever a breakthrough technology appears, capital floods in far faster than the underlying economics can justify. We saw it during the railway explosion of the 1840s.
The same thing happened again during the dotcom era, when companies poured money into building global fibre-optic networks long before demand caught up. Big ideas attract big investment – often too much of it.
And that’s the concern today. The hyperscalers have burned through so much free cash flow that several are now turning to private financing just to maintain their AI spend. For companies that normally mint cash, that’s a warning sign.
Commodity producers
So if AI spending keeps ballooning, where does that leave investors? Personally, I think the smarter opportunities may sit outside Silicon Valley altogether.
If the US and China really are entering an AI arms race — and I believe they are — then the companies producing the minerals that power data centres and the supporting infrastructure could emerge as the real winners.
One commodity already in heavy demand is copper. This year, Glencore (LSE: GLEN) expects to produce around 850,000 tonnes of the red metal. By the end of the decade, that figure is forecast to reach 1m tonnes, with the option to bring on another 1m tonnes relatively quickly if market conditions justify it.
Risks
Naturally, the miner carries risks. Mining is a cyclical industry, and profits can fall quickly if copper or cobalt prices weaken.
The group also has to manage cost pressures, regulatory scrutiny, and the occasional operational hiccup that comes with running large-scale assets. And if the global economy stumbles, even AI-related demand may not prevent short-term price dips.
Bottom line
The AI boom isn’t just about fancy software or sleek robots — it depends on infrastructure, and lots of it.
As generative AI models become more powerful and sophisticated, so too does their energy demand. With an ageing electricity grid, the US is facing a massive upgrade over the next decade, and copper is at the heart of this build-out.
This doesn’t even account for demand from the renewables transition or advanced military hardware in an increasingly polarised world.
Glencore’s scale in copper production positions it well to benefit from this long-term structural demand. By steadily increasing output over the decade, the company could play a pivotal role in supplying the materials that underpin AI growth across the US, China, and beyond.
I expect a bumpy ride as commodity prices can be notoriously volatile. But compared with the valuations of the Magnificent 7, the risk-reward ratio feels very much in my favour. I certainly view the stock as one worthy of further consideration.
