The Digital Oil Fields: Why AI’s Next Trillion-Dollar Opportunity Might Not Be What You Think
The Digital Oil Fields: Why AI’s Next Trillion-Dollar Opportunity Might Not Be What You Think
Published by Willow Rivers Wealth | May 2026
Most investors who want exposure to artificial intelligence head straight for the same short list: Nvidia, Microsoft, maybe a broad tech ETF. They’re not wrong — those are legitimate holdings — but they may be thinking too narrowly. Because if BlackRock CEO Larry Fink is right, we are on the cusp of something far bigger than a semiconductor boom. We may be watching the birth of an entirely new asset class.
And the most interesting parts of it aren’t where most people are looking.
“Futures on Compute”: The Idea That Changes Everything
During a recent public discussion on AI infrastructure and capital markets, Fink made an argument that deserves more attention than it received. He identified four critical shortages that AI is creating right now: compute power, chips, memory, and electricity. These aren’t theoretical bottlenecks — they’re the reason Nvidia’s Blackwell chips had waitlists stretching across multiple quarters, and why Microsoft has openly acknowledged that AI infrastructure constraints have limited cloud growth.
Fink’s insight was this: whenever genuine scarcity emerges in an economically essential resource, financial markets build products around it. Oil did it. Natural gas did it. Electricity did it. Carbon credits are doing it right now. His prediction is that AI computing power — “compute” — is next.
The concept is “futures on compute”: contracts that give companies the right to access a guaranteed amount of AI processing capacity at a fixed price at a future date. Think of it the way airlines lock in jet fuel prices twelve months out to manage their cost base. Instead of barrels of crude, the underlying asset would be GPU-hours, data centre power allocations, or reserved cloud inference capacity.
This is not yet a traded product. But the market is already behaving as though it exists. And that gap between where prices are heading and where most investors are positioned is exactly where opportunity lives.
It’s Not a Software Story. It’s an Infrastructure Story.
Here is the mental shift that matters for investors: the biggest long-term winners from AI may not be the companies writing the software. They may be the companies that own the pipes, the power, the cooling, and the chips that make AI run at all.
Goldman Sachs estimates that global AI-related infrastructure spending could approach $1 trillion over the next several years. Microsoft, Amazon, Alphabet, and Meta are projected to spend over $710 billion combined in capital expenditure this year alone — the majority tied to AI infrastructure.
That spending has to land somewhere physical: data centres that consume enormous quantities of electricity, specialist chips that only a handful of companies can manufacture, networking hardware that moves data at scale, and power infrastructure capable of supplying it all reliably.
The US Energy Information Administration projects that electricity demand from data centres could more than double by 2030. Goldman Sachs estimates AI-related data centres may account for roughly 8% of total US electricity consumption by the end of the decade, compared to around 3% today.
Utility companies — not exactly anyone’s idea of a hot growth sector — suddenly have a legitimate claim to be AI plays. That’s the kind of contrarian reality that tends to reward investors who spot it early.
The European Angle: World-Class Companies in the Infrastructure Stack
The AI infrastructure story is dominated in the media by American names, but some of the most strategically critical companies in the entire chain are European. This matters for investors who want diversified exposure — and it matters because several of these businesses occupy near-monopoly positions in their respective niches.
ASML (Netherlands) — The Irreplaceable Bottleneck
ASML is arguably the single most important company in the global semiconductor supply chain that most people have never heard of. It is the only manufacturer in the world of Extreme Ultraviolet (EUV) lithography machines — the equipment that prints the circuitry onto the most advanced chips. Without ASML machines, there are no leading-edge semiconductors. Without leading-edge semiconductors, there is no advanced AI.
Nvidia’s GPUs, TSMC’s manufacturing process, Intel’s comeback — all of it flows through ASML. As compute demand grows and chipmakers race to build out capacity, ASML is the toll booth at the only road into the future. Its order backlog consistently extends years forward, and its technological lead over any potential competitor is measured in decades.
Schneider Electric (France) — The Power Behind the Data Centre
Every data centre in the world needs power distribution, cooling management, and energy efficiency systems. Schneider Electric is the global leader in all three. Its EcoStruxure platform is the standard infrastructure management system used by hyperscalers and enterprise data centres alike.
As AI drives data centres to consume ever more power, the companies that help them do so efficiently become increasingly valuable. Schneider is not a speculative bet on AI — it is a quietly dominant infrastructure business that happens to be precisely positioned for the decade ahead.
Siemens Energy (Germany) — Electrifying the AI Boom
AI needs power at a scale the grid wasn’t built for. Grid upgrades, transformer capacity, and high-voltage equipment are all in critical short supply globally. Siemens Energy builds the transformers, grid infrastructure, and industrial power systems that will need to be installed at enormous scale to support data centre growth. Recent order books have been running well ahead of production capacity — itself a signal of genuine structural demand.
ABB (Switzerland/Sweden) — Automation and Grid Infrastructure
ABB operates across electrification, automation, and power grid technology. As AI-driven automation accelerates industrial processes and as the grid requires significant investment to handle new loads, ABB sits at a powerful intersection. It is a diversified engineering business, but its grid and electrification divisions are directly in the path of AI-related infrastructure spending.
Infineon Technologies (Germany) — The Chip You’ve Never Heard Of
Infineon doesn’t make AI GPUs, but it makes the power semiconductors that regulate electricity inside the data centres that run them. Power management chips are a mundane-sounding but absolutely essential component of every piece of computing infrastructure. As data centres scale up, so does demand for Infineon’s products. It is a quieter play than the headline names, but one with genuine pricing power in a supply-constrained market.
The Stocks Already Reflecting This Shift
For context, here is how some of the companies most exposed to the AI infrastructure theme are currently valued:
| Company | Ticker | Theme |
|---|---|---|
| ASML | ASML (AMS/NASDAQ) | Chip manufacturing monopoly |
| Schneider Electric | SU (EPA) | Data centre power & cooling |
| Siemens Energy | ENR (ETR) | Grid & power infrastructure |
| ABB | ABBN (SIX) | Electrification & automation |
| Infineon | IFX (ETR) | Power semiconductors |
| Nvidia | NVDA | AI GPU dominance |
| Broadcom | AVGO | AI networking & custom chips |
| Constellation Energy | CEG | Nuclear power for AI demand |
| Vertiv Holdings | VRT | Data centre cooling & power |
| Digital Realty Trust | DLR | Data centre REIT |
What this table illustrates is that the AI infrastructure story spans continents, sectors, and asset types — from REITs to utilities to European industrials. It cannot be captured by a single tech ETF.
The Hidden Story: AI Is an Energy Problem
The most counterintuitive insight in all of this is that AI may turn out to be an energy crisis in technology clothing.
The demand projections are genuinely staggering. Data centres already consume a material share of national electricity grids in the US and Europe. As AI model training and inference scale up — and as AI moves from the cloud into edge devices, manufacturing plants, and autonomous systems — that demand will compound.
This creates an unusual opportunity in energy infrastructure: nuclear, in particular, is experiencing a serious re-evaluation. Constellation Energy in the US has already signed power purchase agreements directly with hyperscalers. In Europe, operators of existing nuclear capacity are quietly becoming strategic assets for tech companies that need large, reliable, carbon-free power sources.
Renewable energy developers, grid operators, and even specialist cooling technology businesses all sit in the path of this spending. The energy story isn’t a footnote to the AI story. For the next decade, it may be the central chapter.
How Willow Rivers Wealth Clients Might Position for This
Every client’s situation is different, and nothing here constitutes personalised financial advice. But for those looking to think constructively about how AI infrastructure themes might fit into a long-term portfolio, here are the frameworks we’d encourage you to consider.
1. Think in Layers, Not Just Headlines
The AI infrastructure stack has multiple layers — each with different risk profiles and return potential. Rather than concentrating in a single headline name, consider whether your portfolio has exposure across: semiconductor manufacturing (ASML), chip design (Nvidia, Broadcom), data centre infrastructure (Vertiv, Digital Realty), power & energy (Schneider, Siemens Energy, Constellation), and networking (ABB, Infineon).
This layered approach means you’re not betting on which application wins, but on the fact that all of them need the same physical infrastructure to run.
2. Use ETFs for Efficient Access to the Theme
For clients who prefer not to hold individual stocks, a range of ETFs now offer targeted exposure to parts of this theme. The iShares Semiconductor ETF (SOXX), the Global X Data Center & Digital Infrastructure ETF (VPN), and European-listed equivalents give diversified access without the concentration risk of individual holdings. For the energy angle, infrastructure-focused ETFs with exposure to utilities and grid companies offer a way to participate in the power demand story.
3. Don’t Ignore European Equities
European investors often underweight European equities in favour of US tech exposure. For the AI infrastructure theme specifically, this may be a mistake. ASML, Schneider Electric, and Siemens Energy are world-class businesses with genuine competitive moats, trading in euros, and with significant exposure to a global infrastructure build-out. For clients concerned about concentration in US dollar assets, these offer meaningful diversification alongside genuine thematic relevance.
4. Consider a Long Time Horizon — and Volatility Tolerance
The AI infrastructure build-out will not be linear. Semiconductor stocks are notoriously cyclical. Energy infrastructure spending can slow with policy shifts. There will be drawdowns, earnings disappointments, and narrative reversals along the way. The strongest case for this theme is a five-to-ten-year horizon, not a quarterly trade. Sizing positions accordingly — meaningful enough to matter, not so large that a correction causes panic — is the discipline that determines whether clients actually benefit from being right.
5. Watch for the New Financial Products
Fink’s “futures on compute” concept is a thesis, not yet a product. But financial markets have a track record of catching up with reality quickly once institutional demand is clear. Over the coming years, we may see structured products, specialist funds, or even direct compute-linked instruments begin to emerge. Clients who have built foundational exposure to the underlying infrastructure theme now will be better placed to evaluate and access those products when they arrive — rather than chasing them after the initial move.
The Takeaway
Larry Fink is not given to empty speculation. When the head of the world’s largest asset manager identifies a potential trillion-dollar asset class in formation, it is worth taking seriously — even if the specific instrument he describes doesn’t exist yet.
The deeper point is this: AI is not just a software revolution. It is a physical infrastructure revolution, and the companies that own the chips, the power, the cooling, and the data centres are building the equivalent of the oil fields and pipelines of the digital economy. Some of the most strategically important of those companies are European, well-established, and currently under-owned by investors who are still looking for AI exposure in purely American tech indices.
At Willow Rivers Wealth, our role is to help clients look a step further than the obvious narrative — and to position portfolios in a way that is thoughtful, diversified, and built to last through the inevitable noise along the way.
If you’d like to discuss how any of these themes relate to your own portfolio, we’d love to hear from you.
This article is for informational purposes only and does not constitute financial advice. Past performance is not a reliable indicator of future results. Always seek independent financial advice tailored to your personal circumstances before making investment decisions.
© Willow Rivers Wealth, 2026