AI Picks and Shovels: How to Invest in the AI Boom (Even If the Bubble Bursts)

Advertising disclosure & disclaimer: This article is for education only and is not financial advice. The companies named are examples of a strategy, not recommendations to buy — always do your own research or talk to a licensed advisor. MoneyPilotAI may earn a commission from links to tools on this page, at no cost to you. How we make money.

Everyone wants to own the company that “wins” AI. The problem: nobody knows who that will be. The chatbot leader of today could be an also-ran in three years, and the most-hyped names already trade at prices that assume near-perfect execution. So how do you invest in the biggest tech wave in a generation without betting the farm on a single winner?

One answer is as old as the California Gold Rush: sell picks and shovels. The miners who chased gold mostly went broke. The people who sold them tools, tents, and jeans got rich either way. Applied to AI, that means buying the companies building the infrastructure every AI player needs — the chips, power, cooling, and concrete — rather than the AI apps themselves. Here’s how the strategy works, the kinds of companies it points to, and the risks nobody should ignore.

Why “picks and shovels” can work even if the AI bubble bursts

Whether or not today’s AI valuations are a bubble, one thing is concrete: the spending is enormous and already committed. The largest cloud and AI providers are projected to spend over $600 billion on capital expenditure in 2026 — up roughly 36% from 2025 — with the bulk of it going to AI infrastructure. That money flows to the same place no matter which chatbot or model “wins”: into chips, data centers, electricity, and the physical guts of computing.

That’s the core idea. A picks-and-shovels company gets paid when the industry builds, not only when a specific AI product succeeds. If the hype cools but the build-out continues — and hundreds of billions in signed commitments suggest it will for years — the infrastructure layer keeps earning. It’s not bulletproof (more on that below), but it’s a way to ride the trend with less dependence on picking the one winner.

The 5 layers of AI infrastructure (with example companies)

Think of AI’s “physical stack” as five layers. For each, here’s what it does and a well-known company that operates there — as an illustration of the category, not a recommendation.

1. The chipmakers and foundries

AI runs on advanced semiconductors, and almost all of the most advanced ones are physically manufactured by a single company: Taiwan Semiconductor (TSM). TSMC holds roughly 72% of the global foundry market and an even larger share of cutting-edge nodes, and it builds the GPUs that power AI — including Nvidia’s. When any AI company needs more compute, the order ultimately runs through foundries like this one. The catch: it’s geographically concentrated in Taiwan, a real geopolitical risk the company is addressing by building fabs in the U.S., Japan, and elsewhere.

2. Data center power and cooling

AI servers run hot — modern AI racks can draw 120–150 kW each, far more than traditional servers — and they need constant, reliable power. Vertiv (VRT) is a leading pure-play here, making the power systems and the advanced liquid cooling that new AI data centers increasingly require by design. As long as data centers get denser and hotter, the cooling-and-power layer stays in demand regardless of which AI model is popular.

3. Precision components and manufacturing

Behind the big names sit specialist manufacturers that build the precise, complex parts the giants can’t make alone. Fabrinet (FN) is one example — a contract manufacturer producing high-precision optical and electronic components used across AI data centers and networking, working closely with major chip and networking companies. It’s a quieter, behind-the-scenes layer, but a critical one.

4. Energy and the electric grid

AI’s appetite for electricity is staggering: data centers could consume roughly 9–10% of North America’s electricity by 2030, up from 3–4% in 2025. That’s driving a once-in-a-generation build-out of power generation and grid equipment — turbines, transformers, switchgear. GE Vernova (GEV) sits squarely in this layer, and its data-center-driven orders have surged. The “AI trade” increasingly looks like an energy trade.

5. Building the data centers themselves

Someone has to physically construct and electrify the buildings. Infrastructure and construction specialists like Sterling Infrastructure (STRL) handle large-scale site development and mission-critical electrical work for data centers. It’s the most literal “picks and shovels” layer — the firms pouring the concrete and running the power for the AI age.

The risks you shouldn’t ignore

“Even if the bubble bursts” is a comforting phrase, but picks-and-shovels investing is not risk-free:

• They’re still tech stocks. If AI sentiment cracks, these names can fall hard alongside everything else, at least in the short term.
• Valuations can already be rich. Some infrastructure winners have soared hundreds of percent; a lot of good news may already be priced in.
• Spending could slow. Today’s capex commitments are huge, but they’re forecasts, not guarantees — a sharp pullback would hit the build-out layer directly.
• Concentration and geopolitics. The chip layer in particular is concentrated in one region, with real political risk.
• It’s not a sure thing. “Less dependent on one winner” is not the same as “safe.” Past performance never guarantees future results.

A common way investors manage this is diversification — spreading across layers, or using a broad technology or infrastructure ETF instead of single stocks. That’s a personal decision; the point is that no single approach removes risk.

How to research these names yourself

If this strategy interests you, the next step isn’t to buy anything — it’s to research. Pull up each company’s chart, earnings trend, and valuation and compare them before you ever consider a position. A charting platform makes this easy: you can screen the sector, track these tickers, and read the technicals side by side. We break down the best research apps in our guide to the best AI investing apps and robo-advisors, and you can chart any of these names free on TradingView.

Frequently asked questions

What does “picks and shovels” mean in investing?

It’s a strategy of investing in the suppliers and infrastructure behind a boom rather than the headline players. The name comes from the Gold Rush, when sellers of mining tools profited more reliably than most miners.

How do you profit from AI even if it’s a bubble?

By owning the infrastructure layer — chips, power, cooling, energy, and construction — that gets paid as the industry builds, regardless of which specific AI product wins. If building continues even as hype cools, that layer keeps earning. It still carries risk if overall spending slows.

Are these five companies recommendations?

No. They’re examples of the five infrastructure layers, included to illustrate the strategy. This article is educational and not financial advice — always do your own research or consult a licensed advisor.

Is there an easier way than picking individual stocks?

Many investors use diversified ETFs that hold a basket of semiconductor, technology, or infrastructure companies, which spreads risk across many names instead of betting on one. Each approach has trade-offs.

Can I lose money with picks-and-shovels stocks?

Yes. They’re still stocks — often volatile tech-linked ones — and can decline sharply, especially if AI sentiment turns or valuations compress. Never invest money you can’t afford to lose.

Why is electricity part of the AI story?

AI data centers consume enormous amounts of power, driving demand for turbines, grid equipment, and new generation. That’s why energy and grid companies have become a core part of the “AI infrastructure” conversation.

---

The bottom line: picks-and-shovels investing is a way to get exposure to the AI build-out without staking everything on guessing the winner — but it’s a strategy, not a guarantee. Understand the layers, respect the risks, and research before you act. See our guide to the best AI investing apps →