What is Validea?
Founded in 2003, Validea provides AI-driven quantitative portfolio solutions and financial research subscriptions for institutional, professional, and self-directed investors.
Based on patented technology, Validea began as an experiment to quantify the investment approaches of the world’s greatest investors using artificial intelligence. Its first such AI “brain,” based on the work of Peter Lynch, saw a 38% total portfolio return in its first year, and a 62% total return in 2009, the year of the Great Financial Crisis stock market crash. Since then, the Validea architecture has grown to twenty-two models, plus a suite of portfolio management and research tools, and newsletter subscriptions.
In 2026, Validea separated its business into an institutional offering, which provides its original suite of models and tools for professional investors and advisors, and an individual offering, which focuses on drawing from its suite to provide solutions and financial research for self-directed individual investors. In its twenty-three-year history, Validea has served hundreds of clients and informed institutional customers with over a billion under management.
Our motto is ex enumeratione veritas, “truth comes in the tally.” Any claim, especially in finance, should withstand mathematical rigor and scrutiny. As of June 2026, all but four of Validea’s twenty-two models have outperformed the S&P 500.
Validea is currently headquartered in the historic Mount Vernon district of Baltimore, Maryland.
The Validea Architecture: From MIT’s Labs to Your Portfolio
Validea owes its architecture to the genius of John Reese. John built his first computer as a pre-teen to help his father, an executive at Capitol Records, with distribution forecasting and logistics. As a teenager he developed his first artificial intelligence algorithm on an IBM 360 owned by the Los Angeles City Department of Water (who allowed John and his friends to use it as night). In a foreshadowing of the “AlphaGo Moment” of 2016 (in which an AI was able to beat the world’s best player at the highly complex, 4,000-year-old game known as Go), Reese built his AI to play four-dimensional tic-tac-toe.
At MIT in the 1970s, Reese worked in the AI labs under mentors and AI pioneers J.C.R. Licklider and Ed Fredkin (it would be Licklider who would eventually suggest Reese apply his AI work to the stock market). While at MIT Reese worked on problems such as linguistics and world modeling – fields that would eventually form the foundation of the today’s AI boom.
After leaving MIT and graduating from Harvard Business School, Reese founded the computing company Micro Networks of America, which he sold in the 1990s to AmeriData (itself eventually acquired by GE Capital). In 1997, he began designing the architecture that would become Validea, earning him two patents in the process. Though semi-retired, Reese still manages his own money using the Validea systems and continues to spend upwards of 40 hours a week on artificial intelligence.
Live vs. “Backtested” Results: A Critical Distinction for Investors
We’ve all seen one – a stock picking system that looks impressive, even bulletproof, until of course you start using it.
Many so-called “backtested” systems are what we call “overfits.” That is, the researcher has created his system to outperform a specific historical sample (the S&P 500 over the last five years, say), buying and selling stocks with uncanny accuracy. He believes the resultant strategy is surely a major discovery. In fact, he has discovered an accident. And an accident that will be almost impossible to repeat…much to the chagrin of any investors who follow it.
At Validea we outperform by doing precisely the opposite. We start with one idea and codify it into a model that can operate autonomously. We then set the model free, giving our subscribers full transparency to judge its results for themselves. Our older models have run live since 2003. Our newer models are a mixture of a live results and hypothetical results based on historical data. We are proud to say that in many cases, as opposed to what is often found in “backested” systems, our live returns have exceeded our hypothetical returns.