AI Model Portfolios: Inside the N.E.W. System

AI model portfolios are the product class the financial newsletter industry converged on in 2026: a quant score, a machine-assembled basket, a quarterly cadence. Whitney Tilson’s N.E.W. System (New Engine of Wealth) at Stansberry Research is a clean specimen, because its mechanics are unusually well documented, on the publisher’s own product pages as well as in the pitch. We read the presentation end to end. Our full AI Civil War teardown covers it in detail. Our The New Engine of Wealth System, Explained explainer covers this in detail.

How the N.E.W. System Is Built

The machine has two layers. First, the Stansberry Score: “we figured out a way of understanding those trillions of combinations and crunching them down into a single number from 0 to 100. The higher the number, the more potential the stock has for growth and vice versa.” The scoring covers 5,300-plus tickers, and the justification is combinatorial: “there are literally trillions of ways to combine these different elements across all stocks listed in America. But AI can do it both instantly and in real time.”

Second, portfolio simulation: “Once we found our highest scoring stocks, the best of the best, we then trained our A.I. to simulate the future outcomes of thousands of different types of portfolios using every combination of companies imaginable.” The output, per the publisher’s own site, is a balanced, diversified 20-stock portfolio drawn from a Russell 1000 scan, delivered as quarterly recommendations plus monthly updates. A position-sizing calculator converts a portfolio amount into per-stock allocations, and Tilson’s role is a human check: “you won’t be blindly following a computer. I still vet each recommendation for you along with a special committee of experts.”

The Evidence Offered

The backtest examples come with scores attached: Cadence Design Systems at 86, Agnico Eagle Mines at 89 (“up as much as 118 percent over the past year”), Telefônica Brasil at 91, Nvidia at 74 “all the way back in 2016,” Sterling Infrastructure at 60-plus since 2022, and the low-score warnings, “Resolve AI, which only scored a 37… would have wiped out 77 percent of your money since the summer of 2024,” and Jet.ai at 15, “wiped out entirely.” The headline backtest: “You’d have outperformed gold, bonds, real estate, and even Berkshire Hathaway. And you’d have turned every $100,000 of your money into $339,000 with less risk than the stock market.”

The free picks demonstrate both directions: Garmin (GRMN) to buy, the fitness-data company scoring “very highly” on its AI subscription monetization, and Rivian (RIVN) to sell, with the financial facts stated plainly (“burning through $2 billion in cash every single year… $5 billion in debt… it keeps issuing new shares”).

The Honest Caveats

Three cautions for the AI-portfolio class as a whole. First, backtests are simulations: the $339,000 figure is “would have” framing, and the presentation itself concedes the genre’s limits nowhere more clearly than in its own disclaimer posture, which points twice to a disclosures page. Second, the “less risk than the stock market” claim for a concentrated 20-stock portfolio is the most aggressive quantitative assertion in the pitch, and the risk measure and period are not defined. Third, factor scoring is legitimate quant practice, but any single proprietary score’s claimed edge is a research claim, not a documented fact; the score-and-outcome pairs shown are, inevitably, the flattering ones.

The scope gap worth noting: the VSL says the system tracks 5,300-plus companies daily, while the product page describes a Russell 1000 scan for the portfolio; the larger figure matches the Stansberry Score screening tool’s coverage. Both are real products; they are different sizes of claim.

What You Get

The N.E.W. System package: quarterly recommendations, monthly updates, full Stansberry Score access, Stansberry GPT (the natural-language screener “powered by hundreds of thousands of dollars worth of real-time financial data”), the Quick Start Guide, “Ripping Point Riches” and “Ripping Point Casualties,” and a mystery bonus. The price is withheld behind the order button, with a discount framed as exclusive to viewers. 30-day full refund, and “you keep everything.” Details in the presentation. For the adjacent pieces, our AI stock screener comparison places the Score in its market, and our Stansberry Research review covers the publisher.

The Honest Read

AI model portfolios are a real advance over the guru-pick genre, and the N.E.W. System is one of the better-documented examples: a defined universe, a disclosed method, a human check, and named free picks with reasoning. The honest questions are the ones every quant product faces: whether the backtest survives live trading, whether the edge persists after costs, and whether a 20-stock portfolio suits your risk. The machine is real. Its edge is the subscription.

NewsletterVetter is an independent publication. We receive compensation from some of the services we review through affiliate links. Nothing on this site is investment advice. Always do your own research.