Fox Business Sep 19, 2026
With Scott Ladner, Chief Investment Officer at Horizon Investments
A data center cannot be built out of semiconductors. It needs land, construction, power, water, roads and the heavy machinery to dig the hole, and Scott Ladner says that is where the second half of the AI capital spending cycle shows up.
The common version of the AI trade is a list of chip names. Ladner's is a chain of suppliers who had nothing to do with technology until the build-out reached them, and who are now looking at orders that did not exist before.
"because this is all new demand for them as well, and new business for themselves"
Scott Ladner, Chief Investment Officer at Horizon Investments, on Fox Business, sets the asset allocation at a firm that runs money for financial advisers, and he came on to name the term he uses for the AI spending cycle. Fox's own description of the segment gives it as a flywheel.
The full segment is covered here so you can skip it.
Here are the 6 insights that matter.
Key Takeaways
The AI capital spending cycle does not stop at semiconductors — the building has to exist before it can hold a chip
The suppliers pulled in are seeing new demand and new business, not a reshuffle of orders they already had
Caterpillar is in the trade because of the heavy machinery needed to dig and build the sites
Copper recently set an all-time high, which he put above $6 a pound in the US and reads as a proxy for how much building is actually happening
The size of the bid over the next five years depends on how many data centers get built, in the US and elsewhere
He is not worried about the doomsday case, on the grounds that regulation takes years to arrive
1. Past the Chips
The segment opened on semiconductor deal-making. The show reported that Intel is in talks with the South Korean memory chip supplier SK Hynix to manufacture chips on US soil for the first time, citing reporting that the arrangement would take in part of Intel's chipmaking site outside Columbus, Ohio, and a joint venture with other large cloud companies looking to lock in memory supply. Intel was the strongest of the semiconductor names on screen at the time, and technology was the only sector working that day.
Ladner was brought on to name the term he uses for the spending cycle as a whole. His answer was that it does not end at the chips, because the building that houses them has to exist first.
Buying semiconductors is the start. After that come the data centers, which have to be constructed, which needs land and infrastructure in place, and which needs people to do the work. The firms that sit at that intersection had no part in the technology trade until the build-out reached them.
2. Diggers, Wires and Water
Asked which areas beyond the picks and shovels he means, Ladner named the parts of the economy a data center has to buy from.
He started with heavy machinery, naming Caterpillar, and with the builders. Then the utilities: the companies responsible for generating electricity and bringing it in, with some of that generation sitting behind the wall at the site rather than arriving off the grid. Then the rest of the physical plant, water as well as electricity, and the roads that have to be built to reach any of it.
The machinery point is the one he tied back to a specific stock. Heavy equipment is what digs and creates the big data centers, which is why he sees Caterpillar trading with the AI complex rather than apart from it.
3. The Copper Tell
The materials question is where the segment produced its one hard number.
Copper has just been at a record
Copper recently had an all-time high
Scott Ladner
He put the level above $6 a pound in the United States. The reason he watches it is quantity rather than price: the metal is a read on how much construction is actually being done, against how much is being announced.
The build-out is a copper story as much as a chip story
we need a lot of copper
Scott Ladner
4. Five Years of Data Centers
The forward-looking part of his case is a volume question. The demand those suppliers are now seeing runs over the next five years, and its size depends on how many data centers actually get built, in the United States and elsewhere.
The investable question is where the money lands on its way through
that is kind of how that cash flow will sort of work its way through the economy
Scott Ladner
He called that one of the most important investment things to figure out over the next few years. It is a different question from which chip company wins: it asks how much of the spending ends up with everyone who supplies a construction site.
5. Not the Doomsday Worry
Ladner said he is not worried about the doomsday naysayers raising concerns about where AI ends up. His reason was procedural rather than technological.
Rules take years, and the years are the opportunity
and regulation takes forever to actually form, as we know that we have some time here, right.
Scott Ladner
The host agreed the existential question is real but said that putting a percentage on it is close to a silly exercise, because nobody can. What he would rather look at is the more likely outcome, which he put as the most transformative technology the world has seen, and what it does to the US and global economy over the next several years.
6. Take a Breath
Asked for his biggest takeaway for the next session after a heavy day, Ladner's first answer was to slow down. The day's move, he said, came down to a couple of things.
One was the language Kevin Warsh had used about the policy decision. The other was speed: moving policy quickly raises both the chance of an accident and the cost of one, because a mistake made at pace is more expensive to come back from.
His closing thought, which the segment cut off partway through, was that the economy is less sensitive now than it used to be.
Bonus Insights
Technology was the only sector working
On the session this was recorded, the program noted that technology was the day's only winning sector, which is the backdrop to a question about whether the AI trade has anywhere else to go.
Why the copper number is the one to hold on to
Of everything named in the segment, copper is the only item with a live public price attached. Caterpillar, the builders and the utilities are positions; the copper price is evidence about whether the build-out that justifies them is happening.
Ladner's bottom line is that the AI capital spending cycle has moved past the semiconductor names into the physical economy that has to build and power the sites, and that the question worth working on for the next few years is how much of that money lands with the suppliers of construction, electricity, water and metal.
Products, Companies & Tools Mentioned
Horizon Investments (His firm, where he sets the asset allocation)
Caterpillar (Named for the heavy machinery needed to dig and build data centers; he sees it trading with the AI complex)
Intel and SK Hynix (The show's lead story: talks to manufacture memory chips on US soil for the first time, reportedly taking in part of Intel's site outside Columbus, Ohio, with other large cloud companies in a joint venture)
Copper (The metal he uses as a proxy for the build-out, recently at an all-time high above $6 a pound in the US)
Listen to the full episode
🔴 YouTube
Watch the full episode:
More Scott Ladner interviews and podcast appearances
If this was worth your time, send it to someone closer to the industry than you are.
Get the latest market chatter as it happens: