Listen to the audio version of this article (generated by artificial intelligence).
Last Sunday night, markets were bracing for the worst.
The price of oil rose while futures fell amid headlines that screamed “Escalation in the Middle East.” In the nutshell, the overall narrative seemed fragile, perhaps even broken…
Then the earnings week began.
What we got was not confirmation of the economy’s weakness.
We’ve got proof that there is Two economies now…And they’re moving in completely different directions.
For months we have been saying that the global economy is quietly splitting in two.
On one side: Companies that are building and supporting the artificial intelligence revolution.
On the other hand: everyone.
In the past week, this idea has started to show up consistently and overwhelmingly in profits.
In order to understand what is happening, we have to zoom out…
Winners: One stacked victory lap after another
The first thing you need to understand is what “AI infrastructure” actually means.
Building an AI-powered data center is more than just buying a bunch of data Nvidia (NVDA) Graphics processing units and their connectivity. It takes a complete physical package: specialized equipment to make chips, memory chips, processors, semiconductors for power management, fiber-optic connections, and — essentially — physical buildings filled with HVAC systems, electrical wiring, and cooling infrastructure.
Each tier of this group reported earnings last week and each tier broke records.
- Chip equipment companies love ASM International (ASMIY) and L Research (LCRX) about high demand and accumulation that extends for years into the future.
- SK Hynix — the dominant supplier of high-bandwidth memory — its revenues rose nearly 200% year over year, with margins more similar to luxury semiconductor stocks.
- Texas Instruments (Texan), the world’s most diversified chip company, saw data center revenue jump 90%.
- Intel (Intech) revealed that AI now drives 60% of its business and is accelerating.
- Visual and communication players like MaxLinear We are seeing triple-digit growth directly related to the expansion of AI clusters.
- Energy and heat infrastructure providers love Vertif (VRT) and J Vernova We are raising guidance as hyperscalers build capabilities at an unprecedented scale.
- At ground level, Comfort systems (It works) — the company that actually builds data centers — is growing revenue more than 50% with a record backlog that continues to rise.
Translation: Buyers no longer negotiate price. They only ask for size. This is a strategic reorientation.
“Customers are prioritizing volume security over pricing, which is maintaining current strength.” — SK Heynckes, CFO, Q1 2026 earnings call
This is a force that is synchronized across the entire industrial ecosystem.
Most importantly, demand exceeds supply… everywhere.
Customers don’t negotiate price anymore.
They fight for size…
Losers: A Guided Tour of the Other Side
Now let’s cross the gap.
Consumer-facing companies and companies not exposed to AI told a very different story.
- Packaging giant Sunoco (son) Reducing guidance as energy and supply chain costs rise.
- United Airlines (firstHe warned that rising oil prices were pressuring margins and weakening demand.
- Service now (now) Longer sales cycles are flagged as institutional buyers hesitate.
- Pegasystems (Bega) fell to new lows after poor results.
- Lululemon (Lulu) replaced its CEO amid slowing growth.
These companies are linked to consumer demand, input costs, and discretionary spending cycles.
Now, these forces are soft.
Why is this structural?
Skeptics will argue that this is just a boom cycle… that AI infrastructure spending will eventually slow, that the winners will concede their gains, and that the divide will narrow as the economy returns to normal. History suggests otherwise, and the earnings data clearly bears that out.
Consider what we heard from management teams across the winning portfolio: SK Hynix said customer demand for memory over the next three years “far exceeds current supply capacity.”
Server CPU demand “has improved over the past 90 days” and momentum extends into 2027, Intel said.
The backlog has already started in 2027, MaxLinear said.
Comfort Systems said its backlog has risen sequentially even as its burn rate has accelerated — meaning it is signing new contracts faster than it can complete existing ones.
These are not companies talking about the hot quarter. These are companies that describe structural supply constraints that are measured in years.
The reason is simple: The physical infrastructure required to run AI at scale — factories, factory equipment, memory capacity, power plants, fiber networks, and data center buildings — takes two to five years from initial investment decision to operational capability. Order here today.
The offer is not.
This gap does not close quickly, which means that pricing power, margin expansion, and earnings growth will last much longer than a traditional cyclical recovery would suggest.
There is also a demand-side structural argument that emerged clearly from Intel and SK Hynix’s texts this week.
AI is evolving from the training phase – where huge models are built in specialized data centers – to the inference and agent phase, where AI actively processes real-world requests at scale, continuously, across millions of concurrent users.
Each agent AI task generates intermediate data that must be stored and processed. Each inference call requires a CPU to coordinate it along with a graphics processing unit (GPU) to compute it. The more AI is deployed, the more is needed from each layer of the infrastructure stack. The demand loop is self-reinforcing.
What this means for your wallet
The bifurcation thesis has now moved from the analytical framework to the documented business reality.
They appear in quarterly revenue pipelines, gross margin ratios, backlog numbers, and earnings per share numbers across the entire AI infrastructure portfolio at one time. This is not a coincidence. It is a structural shift in where economic value is created.
Companies that build, operate, connect, and compute in the age of AI operate in an environment of structurally constrained supply in the face of structurally increasing demand…an economic situation most favorable to sustainable earnings growth.
Out-of-this-world companies operate in an environment where consumer confidence is fragile, input costs are high, and discretionary spending is weak.
The data has spoken. The AI gap will not come….
It’s already here.
If you want the game of picks and shovels to play a role in launching the most ambitious financial products of our generation, you need to See our latest research.
I just released a whole presentation on X money – Elon Musk’s long-awaited attempt to turn X into “Elon’s Bank” — and the public companies that operate it behind the scenes.
I explain the full story on how you can benefit from offering X Money.
