Picks & Shovels: A Smarter Way to Play the Hype

Companies that supply the tools and infrastructure behind a new trend can get paid whichever brand wins.

Finding those companies inside an ETF is harder than it looks. The holdings often tell a different story from the one the fund's name suggests.

The supplier layer of a boom

During the California gold rush, the merchants selling shovels could make money without ever finding gold.

The same idea works in investing. Instead of trying to pick the company that wins a new trend, look for the companies every competitor has to pay.

Take offshore wind. The vessels that install offshore wind turbines are the same class of vessel that installs oil platforms, and a small group of operators serves both industries.

One supplier serves two trends that look like opposites.

Thematic ETFs are where this gets awkward. Morningstar's Global Thematic Funds Landscape (2022) found that fewer than one in 10 thematic funds both survived and beat a broad global equity benchmark over the previous 15 years. More than three-quarters didn't survive at all.

If you want exposure to a theme, look at the companies supplying it before you buy the fund built around the theme itself.

One company, five trends that sound unrelated

GE Vernova is the power and electrification business spun out of General Electric in April 2024.

It sells turbines, generators and grid equipment. It also shows up in ETFs that market themselves as covering completely different things.

Weights come from each fund's most recent N-PORT filing, dated between February and April 2026.

Emerson Electric is the clearest case. It sits in a nuclear fund, a robotics and AI fund, a grid fund and a broad industrials fund at the same time.

The fund names point at very different investments. The holdings say they have more in common than you'd expect.

There's a limit to this. The pattern is much harder to find between traditional energy and clean energy funds: eight traditional energy funds and six clean energy funds share only four companies, all of them renewable fuel producers, and the largest position is 1.52%.

The supplier overlap is real in the industry. The vessel operators are listed in Europe, and US-listed energy funds don't reach them.

The AI chain runs four layers deep

The AI build-out is the clearest version of this, because it buys from the same aisle as the energy build-out.

A data center needs a lot more than chips. Think of the supply chain in four layers:

1. Chips

Processors, from Nvidia, AMD and a handful of others.

2. Data centers

Those chips need buildings, servers and the equipment to run them.

3. Electrical equipment

Transformers, switchgear and cabling to connect all of it to a grid that was never sized for this.

4. Industrial gases

Chip fabs run on high-purity nitrogen, hydrogen, argon and helium.

The table above is where these layers land: GE Vernova and Quanta sell into layer three, Vertiv into layer two.

Every company building AI infrastructure buys from several of these layers. You don't have to guess which AI company wins to own the ones getting paid for the build-out.

The layers are not equally easy to buy

Chips are the easy part: a dozen funds, SOXX at 0.33% and SMH at 0.35%.

Data centers have fewer options, including DTCR at 0.50%. The choice narrows again for electrical equipment, where GRID charges 0.56% and VOLT 0.75%.

You can compare chip and data infrastructure funds side by side.

Then you reach industrial gases, and no US-listed ETF is built around them.

The closest thing is a broader materials fund. State Street's fact sheet put Linde at 12.90% of XLB and Air Products at 4.72% on August 27, 2026. XLB charges 0.08% and nobody sells it as an AI product. Linde plus Air Products is a larger combined position than most theme funds give you in anything.

The further down the supply chain you go, the fewer ETFs you find.

That's useful to know before you go looking for one.

A cheap sector fund may already hold the suppliers

XLI is a broad industrials fund. It has been listed since 1998 and charges 0.08%.

It holds GE Vernova, Vertiv, Quanta Services, Hubbell, Emerson and Caterpillar. GRID holds several of the same names for 0.56%, seven times the fee.

So the check before any thematic buy is whether a 0.08% sector fund already holds the same companies at similar weights.

PAVE is the other comparison worth making.

Global X sells PAVE as a US infrastructure fund, and 16 of its 93 holdings also sit in XLI, together 37% of PAVE by weight. The rest is mid-cap engineering and construction XLI doesn't reach.

The trade-off is cost: 0.47% against 0.08%, or $47 a year against $8 on $10,000. PAVE against XLI puts the two side by side, and expense ratios covers what that gap does over a decade.

XLI is a sector fund and it behaves like one.

The name of an ETF doesn't tell you what you own. Look at the holdings, then compare them with cheaper funds.

If you buy one, know what it adds

A thematic ETF is a satellite position.

Some of its holdings may already be in your portfolio. A total world fund like VT holds 11.47% industrials, and GE Vernova, Vertiv and Quanta all sit inside it at small weights.

Adding a thematic ETF on top raises exposure you may already carry. The question is what the satellite adds to what you already hold.

For most portfolios, the core still comes from broad, low-cost funds bought on a schedule and held for years. A thematic ETF can sit next to that core if you want a more targeted position.

The supplier layer is a sturdier way to hold a theme. It's still a theme, bought after the story was already told.