The entire conversation around investing in artificial intelligence has been almost comically dominated by one name: Nvidia. It has become the default answer, the pub quiz certainty, the stock-market equivalent of saying “The Beatles” when asked to name a famous band. And for good reason, its dominance is undeniable. But what if this hyper-focus is causing us to miss the bigger, and perhaps more lucrative, picture?
While the world understandably fawns over the star quarterbacks designing the graphics processing units (GPUs), a quieter, more strategic game is unfolding in the trenches. This is the world of AI infrastructure investments—the plumbing, the wiring, and the physical foundations upon which the entire AI revolution is being built. It’s less glamorous, sure, but it might just be the smarter long-term play.
The Real Guts of AI
When we talk about AI infrastructure, what are we actually talking about? It isn’t just a cloud of clever algorithms. Think of it like building a new city. You can’t have gleaming skyscrapers (the AI models) without the concrete, steel, power grids, and road networks (the infrastructure).
This infrastructure has three core components:
– Data Centres: The physical buildings that house everything, complete with colossal power and cooling systems.
– Specialised AI Hardware: This is more than just Nvidia’s GPUs. It’s the high-bandwidth memory needed to feed those processors, the networking equipment that allows thousands of chips to talk to each other, and the racks that hold it all together.
– Software and Networks: The connective tissue that orchestrates the hardware and manages the immense flow of data.
For a while, the market was happy to throw money at “story stocks”—companies with a compelling AI narrative but perhaps not the balance sheet to back it up. That era is drawing to a close. The new focus, as highlighted by savvy investors, is on tangible business impact. The question is no longer “do you have an AI story?” but “how much are you earning from the AI build-out?”.
Following the Money, Not the Hype
This is precisely the philosophy of Michael Seidenberg, manager of the £1.9bn Allianz Technology Trust. As detailed in a recent TrustNet article, his strategy is refreshingly pragmatic. Instead of chasing every shiny new AI application, he’s focused on identifying the companies that are indispensable to the physical construction of AI.
The results speak for themselves. The trust has delivered a staggering 72.8% return over the past five years, more than doubling the sector average of 35.9%. How? By looking past the obvious and betting on the enablers.
The Unsung Heroes of the Chip Boom
Nvidia’s GPUs are the brains of the operation, but brains are useless without a body. That’s where semiconductor alternatives and other crucial hardware come into play. Seidenberg’s portfolio gives us a masterclass in this approach.
Take Micron Technology, a company the trust has held since 2016. Micron doesn’t make the glamorous GPUs; it makes memory. Specifically, the high-bandwidth memory (HBM) that is absolutely critical for training and running large language models. As AI models become larger and more complex, their appetite for memory is insatiable. This makes Micron a classic “picks and shovels” play in the AI gold rush. The result? A 313% surge in its share price over five years.
Then there’s the even more obscure, yet wildly successful, pick: Celestica. You could be forgiven for asking, “Who?” This Canadian company provides design, manufacturing, and supply chain solutions for the tech industry. They are a key part of the ecosystem that assembles the complex racks and systems that fill modern data centres. They are, quite literally, building the physical form of AI. Their reward has been a mind-boggling 3,561% share price increase over the same five-year period. This is the power of investing in the essential, but often overlooked, cogs in the machine.
Building a Resilient Portfolio
The beauty of Seidenberg’s strategy, and a lesson for all of us, is that it’s not about being anti-Nvidia or avoiding the Magnificent Seven. In fact, five of them are in his top holdings. It’s about creating breadth and resilience. It’s about understanding that the AI ecosystem is vast and interconnected.
By investing in tech trust portfolios like the one managed by Allianz, investors gain access to a diversified basket of these companies. These trusts offer professional management that can identify the Microns and Celesticas of the world, while also maintaining strategic positions in the established giants. What’s more, some of these trusts, including Allianz Technology Trust, currently trade at a discount to their Net Asset Value (NAV)—in this case, around 7%—meaning you can effectively buy into this expertly curated portfolio for less than the market value of its individual holdings.
Of course, not every bet pays off. The trust’s biggest detractor was the software company Atlassian, whose share price fell 41%. This, in itself, is telling. It reinforces the idea that in this current phase of the AI boom, the market is rewarding the tangible builders of specialised AI hardware and infrastructure more than the software players whose AI-driven profits are still more of a promise than a reality.
The Long Game for AI Infrastructure Investments
So, is this all a bubble? Seidenberg thinks not. “I’m not worried about AI being a bubble long term,” he states. “The reality is that it has huge potential to disrupt and influence many industries.”
This is the critical takeaway. The infrastructure being feverishly built today is not just for running chatbots. It is the foundation for a complete industrial transformation. It will power AI-driven drug discovery in pharmaceuticals, high-frequency trading in finance, autonomous logistics in retail, and predictive maintenance in manufacturing.
The demand for the underlying plumbing—the memory, the networking, the cooling, the power—is set to have a long and durable tailwind that will likely outlast the hype cycles of individual applications. While we will certainly see winners and losers in the AI software race, the companies laying the tracks will almost all benefit from the increased traffic.
Ultimately, the story of AI infrastructure investments is a reminder to look at the whole system, not just the most visible part. The gold rush was profitable for some prospectors, but it was dependably profitable for the people selling the picks, shovels, and denim jeans. In the great AI build-out of the 21st century, the same logic applies.
What are your thoughts? Are you focused on the big-name players, or are you digging deeper for the hidden infrastructure gems? Let me know in the comments below.
—
Further Reading
– For a deeper dive into Michael Seidenberg’s strategy, read the full analysis on TrustNet.
– Follow the latest trends in tech trust portfolios and semiconductor markets to stay ahead of the curve.
