Philosophically, I don’t believe in writing thematic posts and dumping a basket of a hundred stocks on my readers, then claiming victory when a few perform. To me, that would be very disingenuous. Even when I have done thematic posts, I have limited the number of stocks to just a few. I very well understand that people read this blog to generate alpha, not for amusement.
For the first time, I am going to do a small basket for investors looking to build a portfolio that will benefit from the AI revolution over the long term. I have already written about the majority of the names included in the portfolio. This is for investors who want broad exposure and want to “set it and forget it.”
This will be short and sweet, but absolutely actionable. After all, our tagline is: Simple. Incisive. Actionable.
I wrote three articles in June covering the AI ecosystem in detail. A friend called them the “big money” articles - home runs. I humbly request that you read them if you want detailed insight into why the names in the portfolio made it.
Networking
High-speed Ethernet switches, routing systems, and AI fabric architectures are required to connect thousands of power-hungry GPUs without creating data bottlenecks.
Stock: ANET.
Servers
The datacenter ramp-up continues, albeit with hiccups. Corporates have started ramping up AI server procurement in earnest. After the ramp-up, the regular cycle of upgrades and replacement will continue.
Stock: DELL.
Cybersecurity and Infrastructure Monitoring
As I write this article, FinX is abuzz with news of Frontier Labs hacking thousands of websites, including government ones. Well, since I used to be an Enterprise Security Architect for a Fortune 150 company from 2002 to 2005, I decided to chime in:
This is like locking up your house and leaving the keys hanging on the front door. As I said earlier, this is total amateur hour all around.
OpenAI needs to hire competent infrastructure and network engineers.
Those getting hacked need to get their own act together.
In a way, what is happening is good. It is exposing vulnerabilities and incompetencies across the ecosystem. Since cybersecurity and infrastructure monitoring are now on people’s radar, the issues will get addressed. There will be noise for a few days, then people will move on.
Stocks: PANW, NET, DDOG.
Data
As I wrote on June 15:
Why do you think a company like Snowflake (SNOW) even exists? They shine at enabling organizations to store, manage, and analyze massive amounts of data in a single, unified system from siloed data sources.
A unified data view is important for enabling AI implementation in enterprises. Currently, the majority of enterprises are, to varying degrees, getting their internal systems ready for AI. Part of this effort is to fix their data.
Stock: SNOW.
SaaS and Custom Software
In my June 1 article, “AI Ecosystem Update”, I had a section: “AI in SaaS and Enterprise IT”. See how I worded it? “AI in …” Here is how I explained it in the follow-up June 15 article:
While Anthropic talks a good enterprise game, all of the enterprise solutions currently in production are mostly Anthropic’s LLM running on Palantir (PLTR). People pontificating that Frontier Labs will somehow make PLTR redundant don’t understand what PLTR does technically. They don’t understand ontology, and that PLTR is like an embedded operating system, the orchestrator. LLMs don’t have these capabilities.
Why do you think the Frontier Labs need Forward Deployed Engineers or Solutions Consultants? They need these people to go into every enterprise, build the capacity, enhance capabilities, and provide the solution. That is easier said than done. Getting into a Fortune 500, building trust, and then being given the keys to mission-critical systems takes a lot of time. The faster path to enterprise adoption is for Frontier Labs to piggyback on consulting companies or companies like PLTR that are already embedded in the Enterprise IT ecosystem.
When I wrote this, PLTR and ACN (Accenture) were getting hammered. Guess what? They have rebounded nicely since then and then some. Who was going to be right? A former high-level Fortune 500 IT executive (I moved on from IT to other things a decade ago) or Wall Street finance bros who have no clue how SaaS and Enterprise IT work?
Stocks: PLTR, ACN, NOW.
Water
Equity markets are hyper-focused on the power grid bottleneck for Hyperscale AI data centers (e.g., nuclear power purchase agreements, natural gas turbines, grid interconnection queues). The market assumes water is a simple municipal utility input that scales linearly via cooling towers.
Here is what I wrote on August 8 about AI data centers and water:
Water is a far more localized, politically volatile, and physically constrained bottleneck than power.
The Micro-Geographic Mismatch: Power can be transmitted across regional ISOs (Independent System Operators); water cannot be moved easily over long distances. AI clusters are concentrated in dry or water-stressed regions (e.g., Phoenix, Northern Virginia, Dublin).
Evaporative Consumption vs. Withdrawal: Traditional evaporative and adiabatic cooling towers lose 100% of consumed water to the atmosphere as vapor. Hyperscalers consume billions of gallons of municipal drinking water per day globally.
Local NIMBYism & Regulatory stranded-asset risk: Municipalities are increasingly capping data center water allocations to protect municipal residential tariffs. A $2B AI compute facility can be rendered inoperable not by power outages, but by municipal water rationing or local environmental injunctions.
Companies that provide solutions across the entire water lifecycle, from extraction, treatment, and distribution to testing, monitoring, and wastewater reuse, will benefit from this.
Stocks: ECL, XYL.
Electricity
On July 23, 2025, I wrote the following on X:
There is a lot of talk about electricity requirements for AI related data centers. The following is not being discussed much:
Upgrading the grid and ensuring grid stability
Water availability and management
In July, New York halted large data center construction over grid concerns. Now, Texas has imposed a moratorium pending grid and environmental review. I asked the following sarcastically on X:
What happens to the AI trade if there is a nationwide moratorium, like Texas's, to review the grid? Any engineer worth his salt would have told you the grid needs upgrading to handle so much demand. How much time will it take to scale up?
In any event, the grid will need upgrading and ongoing maintenance.
Stock: PWR.
Corporate Liability Insurance
How can I write an article without a curveball? Something "straight out of left field"? Well, here it is.
FTC Chair said a few days ago that humans behind AI agents should be held criminally responsible for their behavior. Oh, this is going to be an interesting debate. From an investment perspective, corporate and commercial insurance brokers are going to make hay. Everyone will need liability insurance to cover AI missteps on top of what they already have.
Stock: MRSH.
Portfolio
ANET
DELL
PANW
NET
DDOG
SNOW
PLTR
ACN
NOW
ECL
XYL
PWR
MRSH
No semis? I am not interested in shovels; I want the gold. The other thing about this portfolio is that even if, for whatever reason, the semis crash, these companies will hold up, as they are not totally dependent on AI, and when the tide turns, they will be big winners.
Note
This is just my opinion, and I am no one to tell you how to trade or invest. How you go about building the portfolio is totally up to you.
Quick Updates
Coal: In May, I wrote about how coal was going through a mini-rejuvenation due to AI-driven electricity demand. We got a nice surprise over the weekend: China has agreed to buy 10 million tons of US coal in 2027 and 2028. This news should act as a tailwind for the coal sector.
Diesel: A day after I published my medium- and long-term view on diesel, the White House reversed its position on banning diesel exports. When I published, the White House was not considering a diesel export ban. The situation is still in flux. Any talk or action on diesel exports will induce volatility in oil refinery companies (MPC, VLO, PSX). Irrespective of how this plays out, my views, as laid out in my article, remain unchanged.
