Everyone is hunting for the next AI winner.
The obvious names have already been found. Nvidia, Microsoft, Amazon, the hyperscalers and infrastructure providers. The market has spent two years chasing anything with an AI label attached to it and are now borrowing deep for cheap companies in the AI chain.
But the interesting opportunities can be many steps further down the chain and non-obvious at first look.
I've been thinking about what happens as AI spending moves from billions to trillions.
The easiest way to understand it is through something mundane. Imagine you're a capacitor manufacturer. Suddenly demand doubles because every server, data center and AI system needs more electronics. Do you immediately build a giant new factory? Probably not. History teaches manufacturers that today's shortage often becomes tomorrow's glut. Instead, they run harder, increase prices and harvest margins.
We've already seen versions of this elsewhere. Dell didn't invent AI, but it sells the boxes AI runs on. IBM's opportunity isn't AI itself but trust. Governments and large corporations facing complex technology choices often buy reassurance as much as hardware or software. IBM has been selling reassurance for decades. Suddenly, up these old dogs of yester-year go, because who can feel the AI beast at scale. Not some little plucky startup, instead it will be the big beasts.
The same extended logic points to my favour offbeat beneficiary: Goldman Sachs.
The AI buildout is becoming one of the largest capital spending programs in history. Data centers, power infrastructure, semiconductors, transmission networks and industrial reshoring projects all require staggering amounts of financing.
Someone has to arrange that money.
As hyper-scalers and infrastructure developers raise untold billions of dollars and soon trillions, investment banks go from essential, existential. Vast amounts of new debt must be structured, syndicated, packaged and distributed. New equity must be raised. Acquisitions financed. Balance sheets managed. AI bubble = investment bank boom.
Goldman is not an AI company at all. It is a financial infrastructure company sitting at the center of the AI investment cycle that anchors the whole process.
No finance, no AI.
The broader reindustrialisation of America only strengthens the case. New factories, energy projects, grid upgrades and strategic supply chains all require capital. If Washington is serious about rebuilding industrial capacity, vast amounts of new money will need to flow through the financial system.
Historically, investment banks have done exceptionally well during major investment booms. Railways, telecoms, the internet and private equity all generated enormous fee pools for the firms moving capital from investors to builders.
AI will prove more so.
At roughly 17 times earnings and paying a dividend, Goldman does not carry the valuation normally associated with AI-linked stocks. The market largely sees it as a traditional financial institution rather than a participant in the AI buildout. That because the market hasn’t research far enough down the linkage to understand that wall street will be a critical to the AI bubble as any memory maker.
In the AI era, investment banks like Goldman Sachs will be financing the entire expedition and as is traditional making out like bandits.
17 P/E….1.75 dividend. Need you know more?