Porter Stansberry has a new portfolio. Five companies. Each one is a bottleneck in the AI supply chain. He calls them the chokepoint stocks — the essential infrastructure that AI cannot function without. The question is whether they are genuinely irreplaceable toll roads or just good marketing wrapped around a familiar list.
The Silicon Dollar macro thesis — the argument that AI infrastructure is replacing oil as the dollar’s global demand anchor — has been covered separately. This is about the stocks themselves. What they are, why they are supposed to be chokepoints, and which ones are actually new ideas versus recycled favorites from previous Porter promos.
The Chokepoint Logic
The framework is simple and it is not original to Stansberry. Keith Kaplan at TradeSmith pitches a similar idea with different picks. The concept is that in any large-scale technological transformation, the bulk of the value accrues to the companies that control the physical bottlenecks the technology cannot bypass. The companies that own the scarce infrastructure. The toll roads.
Stansberry’s version identifies five bottlenecks in the AI physical stack. Power delivery for the chips that train the models. The metals that go into every semiconductor and data center connection. The natural gas that runs the 24/7 data centers. The mineral rights under the Permian Basin, where the cheapest gas in North America meets the AI buildout. And the water infrastructure that keeps Permian production — and therefore that cheap gas — from grinding to a halt.
Each one is a dependency. AI cannot grow without them. The question is how much of that dependency is already priced in.
The Power Delivery Chokepoint
The most interesting pick in the portfolio is a Massachusetts-based power electronics firm. It builds modules that convert raw electricity into the ultra-high-current, low-voltage power that Nvidia’s most advanced GPUs require. The problem is physical. Next-generation AI racks are liquid-cooled and densely packed. Conventional voltage regulators from Monolithic Power Systems, Renesas, and Infineon physically do not fit. This company’s patented architecture does.
The stock has already run. It traded around $50 in August 2025 and hit roughly $275 by May 2026 — a 5x move in nine months. The catalyst was real. An ITC ruling in February 2025 permanently banned imports from Foxconn, Delta Electronics, and three other Asian rivals. The company has 800 patents it is actively enforcing. It manufactures 100% domestically, which means it is also protected by the 100% tariffs on Chinese competitors.
The financials back the story up. Backlog grew 70% in a single quarter to over $300 million. Revenue hit $113 million in Q1 2026, up 20% year over year. Licensing revenue went from $2.8 million in 2022 to a $15 million quarterly run rate. Expecting roughly $300 million in settlements through 2026.
Stansberry claims he has never discussed this company publicly before. That is almost certainly true — it is not a household name. But the stock has already priced in a lot of the good news. At a $12 billion market cap, the easy money may have been made. The question is whether the patent moat and the AI buildout can sustain the growth trajectory from here.
The Metals Royalty Chokepoint
The youngest company in the portfolio is four years old. It follows the Franco-Nevada model — pure royalty streams on mining assets, no operating expenses, no employees at the mines. Just mailbox money from every ounce of silver, copper, and gold produced.
The thesis is straightforward. Silver goes into every semiconductor. Copper is in every wire and connector. Gold is in every contact point. AI data centers and chips consume enormous quantities of all three. This company owns royalty streams across four continents that collect on that production.
The financials are early-stage. Revenue was $12 million in 2024. The company forecasts roughly $100 million in 2026, targeting $300 million by 2028. The jump is ambitious. The royalty model is capital-light and scales well when it works, but the company is pre-proving at this stage.
Stansberry compares it to Franco-Nevada, which he recommended nearly 20 years ago and which has compounded extraordinarily since its 2007 IPO. The comparison is fair as a business model parallel. As a valuation parallel, it is aspirational. This is a $12 million revenue company projecting $100 million. Franco-Nevada does billions. The royalty model is proven. This specific royalty company is not.
The Natural Gas Chokepoint
America’s dominant natural gas producer generates roughly 6% of all U.S. natural gas at the lowest cost in the country. In a single month, it produced $1 billion in free cash flow. Forward PE is around 12x.
The AI thesis is that data centers need 24/7 baseload power. Solar and wind cannot do that. Nuclear takes a decade to permit. Natural gas is the bridge fuel, and this is the lowest-cost producer. The thesis is logical. The problem is that this is also the most recycled pick in the portfolio.
Stansberry has been pitching “Gods of Gas” — this same company as the premier AI energy play — since Porter & Co. launched in May 2022. It has appeared in the 1776 Moment promo, the 2026 AI Playbook, the Final Melt-Up All-In Playbook, and now the Silicon Dollar. It is a good company with a strong turnaround story — the Rice brothers took control in a 2019 proxy battle and grew free cash flow from negative $3.4 billion to over $7 billion. But it is not a new idea. Anyone who has been following Porter for more than a year has already heard this pitch.
The forward PE is reasonable at 12x. The cash flow is real. The AI data center demand is real. But natural gas is a cyclical commodity, and this is a commodity producer, not a toll road. The chokepoint thesis works better for companies that have structural moats, not price exposure.
The Permian Duo — Mineral Rights and Water
These two stocks are best understood together. One owns 87,000 mineral acres in the Permian Basin. It collects royalties on the oil and gas produced from those acres. No employees. No rigs. No operating costs. Pure mailbox money.
The other owns the largest produced water pipeline network in America, also in the Permian. Produced water is the wastewater that comes out of the ground during oil and gas extraction. The Permian generates enormous volumes of it. If there is nowhere to put it, production stops. This company built the solution.
The AI connection runs through cheap natural gas. The Permian Basin sits on top of the cheapest natural gas in North America. AI data centers are being built there because of it. As AI buildout accelerates, energy demand on the mineral rights holder’s land rises, and royalties follow automatically — zero additional spend. The water company collects its toll on the water side. When regulators shut down competing water disposal options, its market share in the Delaware Basin jumped from 5% to 39% in a single year. It is now part of a West Texas duopoly with roughly 50% EBITDA margins.
The Permian supplies roughly 50% of U.S. domestic oil output. The thesis is that there is practically no Permian production without the water infrastructure company, and no cheap gas for AI data centers without Permian production. The dependency chain is real. But the mineral rights holder is also an energy royalty company — it is exposed to commodity prices, not just data center buildout. And the water company is priced like a boring pipeline business, which is exactly what makes it interesting if the AI demand thesis plays out.
Fresh Picks and Familiar Faces
The honest assessment of this portfolio is that it is a mix. The power delivery pick is genuinely new. It has a real patent moat, a real ITC ruling, real domestic manufacturing, and real revenue growth. It is also up 5x in nine months, which means the easy entry point is gone.
The metals royalty pick is early-stage and ambitious. The royalty model is proven, but the company is pre-scale. It is a bet on execution, not just thesis.
The natural gas giant is a good company and a recycled pitch. It has been in every Porter promo since 2022. The AI data center demand thesis is real, but the stock is not a discovery. It is a commodity producer with a cyclical price.
The Permian duo — mineral rights and water infrastructure — are the most interesting of the energy plays. The Permian AI data center buildout is happening. The royalty model and the water monopoly both have structural advantages. The question is how much of the AI demand is priced in versus how much is still speculative.
What is missing from the portfolio is notable. There is no semiconductor company. No hyperscaler. No data center REIT. The chokepoint thesis explicitly excludes the obvious AI winners — Nvidia, Microsoft, Amazon — and focuses on the infrastructure they depend on. That is the strategy. Own the toll roads, not the traffic. It is a coherent framework. Whether it produces better returns than owning Nvidia depends entirely on whether the toll roads compound faster than the traffic.
What This Portfolio Actually Is
Stansberry’s Silicon Dollar portfolio breaks down into three categories. One genuinely new idea with a strong moat that has already run. One early-stage royalty bet. And three energy plays that are variations on themes he has been pitching for years.
The chokepoint framework is worth understanding. The idea that AI infrastructure creates structural bottlenecks that are investable — that is a real insight, and it is supported by policy moves, capital flows, and institutional validation from the CME, the State Department, and the world’s largest hyperscalers. The specific stocks are a mixed bag of execution risk, cyclical exposure, and timing.
The investor who buys this portfolio today is betting that the power delivery company’s patent moat sustains its growth, that the metals royalty play scales its model, and that the Permian energy thesis plays out through the mineral rights and water infrastructure picks. Those are three separate bets with three different risk profiles. The chokepoint framing makes them look like a single thesis. They are not. They are five companies at different stages, with different moats, and different levels of exposure to the same macro trend.
The macro trend is real. The execution is what will separate them. More Porter Stansberry files are collected in the promo watch hub.