I spent years on Wall Street building the same models that gave big institutions their edge, watching the gaps and building the signals to close them. It became impossible to ignore how the tools that made investors their money stayed locked behind institutional walls, never reaching the individual investors on the other side of the trade. Eventually I’d seen enough of the game from the inside. I left the big institutions to share what I’d learned with the people who never had that access: retail investors with just as much of a right to win.
In 2019, that desire finally manifested as Prospero.ai, an AI app designed to simplify institutional-grade data into signals everyday investors could use, with accuracy that would earn their trust over time. Today, what started as an app is now an ecosystem that democratizes access to institutional intelligence, helping so-called “dumb money” get smart.
In the years since, much has changed. Retail traders are no longer at the margins. In fact, in 2025, retail traders accounted for 20–25% of total volume, reaching a record high of 35% in April that year, per J.P. Morgan data. What’s even more interesting is that they’re coming out on top: An analysis of retail investing patterns in 2025 showed that individual investors outperformed both SPY and QQQ, two of the most widely-held professionally managed index funds, designed to mimic S&P 500 and Nasdaq 100 returns.
The professionals have noticed, and the relationship has shifted. According to a recent research report from fintech platform Exabel, 74% of institutional investors began using alternative data within the past five years, and 93% are expecting an increase in their use of social sentiment data specifically over the next three years. This means that the investors earning banner returns aren’t just reading earnings reports and listening to analyst calls anymore. They’re tracking Reddit threads alongside institutional reads like options flow and dark pool activity, identifying signals that show up in the data long before markets react—if you know where to look. What used to look like a handful of forces moving markets has multiplied into dozens, most of them unseen until someone refines the raw data into signal. Algorithmic trading rose to handle exactly that, but compute alone isn’t enough to maintain an edge.
Institutional players have spent the last decade or so rebuilding themselves around code, and the trend shows no signs of stopping. In fact, algorithmic and high-frequency strategies now account for at least half of U.S. equity trading volume, as of 2024. Even the most prominent skeptics have given in. Cliff Asness, co-founder of AQR Capital Management and a longtime public critic of overreliance on machine learning, told the Financial Times his firm has “surrendered more to the machines” after years of holding the line.
But there’s a catch. While all funds aim to build the best technology to maximize returns, high competition and headline risk often lead to homogenous thinking. Funds constantly poach talent from one another, so a small circle of people converges on the same investment strategies, computational approaches, and data sources, which ultimately radiates across the entire industry.
I knew that pattern firsthand, which is why I didn’t try to maneuver around it. When I founded Prospero.ai, the goal wasn’t building market-beating signals. The community we built by sharing our signals became a feedback loop: a group of more than 200,000 monthly users who put our signals to work, saw real returns, and grew more confident with each trade, providing us with behavioral data that then sharpened the platform’s signals over time. The eventual result was a happy accident. In building the loop, we had created a genuine edge and given ourselves a new and highly complex way to compress and test massive amounts of data.
The edge has shown up in two places at once: in the community and in the public markets. What I wrote about in Fast Company earlier this year keeps happening in real time: A useful read reaches retail investors. They share it and build conviction around it. They put capital behind it, sometimes within hours, making the initial signals even more powerful.
My own public appearances have borne this out. Remarks on Prospero.ai’s top picks, based on our signal research, reached retail investors. Then these comments drove them to act and the markets responded with steep climbs within a month. The relationship may not be one-to-one, but as I saw the signals work and the public trusting them (which made them even more accurate), a natural question arose: How could we best expand this model?
Aethon Fund became the answer. It’s a way to pour gasoline on the flywheel I inadvertently developed in founding Prospero.ai, which could only build signals as fast as its own resources allowed. It became clear that a hedge fund deploying these signal libraries at institutional scale would grow the ecosystem and capitalize on its ideas better, building out new signals that will get shared with the public where useful. The result is a symbiotic relationship that ultimately allows us to meet people earlier in their investing journeys and enhance how we teach them to use our signals, continuing the mission that started this journey.
George Kailas is Founder & CEO of Aethon Fund and Chairman of Prospero.ai.