The CNC Group
Free report · CNC Group Research · 2026 outlook

The next $100 trillion is infrastructure.

AI is not an app cycle. It is an infrastructure supercycle — and like railroads, electrification and the internet before it, the money goes to whoever owns the physical layer. This report names the layer: Compute, Data and Energy, and seven companies positioned at the centre of it.

By Armando Pantoja13 pages · PDF · free7 stock picksPublished 2026

Get the report — free

No charge, now or later. Put in your email and the PDF opens straight away — no drip, no sales sequence. One email, the report.

We keep your address to send you future research. Unsubscribe any time.

The thesis

Every model needs four things. Three of them are investable.

Training and running AI requires large data centres, advanced chips, high-speed networks and massive amounts of electricity. Software captured the last cycle's value. This one shifts it back to the people who own the hardware, the data and the power.

01

Compute

The chips and hardware that process AI workloads — and the one company whose machines every advanced chip on earth is made on.

02

Data

The fuel that trains the models and gives them decision-making capability — and who is already paid to organise it for governments and enterprises.

03

Energy

The hidden bottleneck. Data centres need power that is reliable, dense and constant — which is why nuclear is back in the conversation.

What's insideSeven picks · reasoning shown

Seven companies. One framework.

Not a list of tickers — each name is placed in the Power Triad and argued for: what it controls, why that control gets more valuable as AI scales, and what would make the thesis wrong. One is below; the rest are in the report.

Compute infrastructure
NVIDIAand the bottleneck

Why demand for AI compute is outrunning supply — and who else benefits.

Compute · the machines behind the chips
ASML

The company every advanced semiconductor on earth depends on.

Data infrastructure
Palantir

Turning raw data into decisions for governments and the Fortune 500.

Energy · nuclear
Oklo & Cameco

Small modular reactors and the uranium that feeds them.

Security for the AI era
CrowdStrike

Every new model is a new attack surface.

Plus one more
Another name

The seventh pick and the reasoning behind all of them.

0.6%a year added to global productivity by generative AI through 2040 — McKinsey's estimate, cited in the report
4things every model needs: data centres, chips, networks, electricity
13pages — written to be read in one sitting, not filed
The author
Armando Pantoja

Armando Pantoja

Financial futurist and USA Today bestselling author (The Strategic Millionaire; The Future of Wealth, 2026). Fifteen years in financial technology, a TEDx speaker, and the host of The Future Wealth Conference. Every public call he makes is on the record — which is the only reason to read a report like this one.

CNBCYahoo FinanceNYSEABC

The report is the thesis. The CNC is where it gets worked.

Armando's private research desk and investor community — live with him every Monday, every call on the public record, and the full reasoning behind every position.

Join the CNC

Educational and informational only. Nothing on this page or in the report is personalised investment advice, an offer, or a recommendation to buy or sell any security. Armando Pantoja and members of the CNC may hold positions in companies discussed. Past performance is not indicative of future results. Do your own research.