Can an AI really beat the market? That’s what Nathan Smith, a high school student from Oklahoma, set out to discover. Armed with €100 and a bit of curiosity, he handed full control of his investment strategy to ChatGPT—and within just four weeks, his tiny portfolio was outperforming the market by a wide margin.
The results were impressive. But before you start dreaming of AI-powered riches, let’s break down what really happened.
It all started with curiosity
Nathan had seen plenty of ads for so-called AI stock pickers, promising smarter, faster returns. But he wondered—what if he skipped the middlemen entirely and just asked a language model to build a portfolio from scratch?
His brief was simple: create an all-stock portfolio made up exclusively of small-cap U.S. companies, each with a market cap under $300 million. ChatGPT accepted the challenge and returned a fully diversified basket of picks.
Fast-forward four weeks: the portfolio had jumped by 23.8%, while the Russell 2000, the benchmark for small-cap stocks, had barely nudged past 3.9%.
The AI called all the shots
What set this experiment apart wasn’t just the gain, but the process. Nathan didn’t make a single stock decision himself. ChatGPT chose everything—from position sizing and stop-loss levels to weekly rebalancing calls.
Interestingly, the chatbot even showed signs of caution that many human traders might lack. One standout moment? It recommended selling a stock that had gained 50%, citing the high volatility typical of micro-cap assets. A calculated move that turned out to be wise.
Nathan’s only job was to manually execute the trades and double-check for the occasional conflicting instruction. Otherwise, the AI ran the show.
A promising start, but far from conclusive
While the short-term numbers are eye-catching, a four-week performance spike isn’t exactly proof of a long-term investment strategy. Financial markets are notoriously unpredictable, and small-cap stocks can swing wildly based on factors that even AI can’t anticipate.
Regulators like the AMF frequently remind investors that strong short-term results don't guarantee future performance. Nathan knows this, too. That’s why he’s planning to extend the test over a full year before making any bold claims.
In the meantime, the experiment has sparked a deeper interest in quantitative finance and coding—two skills that could serve him well in or outside the market.
The fine print investors shouldn't ignore
Stories like Nathan’s are inspiring, but they also come with clear caveats. Leaving your money in the hands of an algorithm, especially without a strong understanding of how the system works, can be risky. Micro-cap stocks, in particular, are known for sharp price swings and low liquidity, making them both exciting and dangerous.
Yes, ChatGPT built a portfolio that surged in the short term. But expecting that kind of success to repeat without any bumps along the way? That’s more wishful thinking than sound investing.
Bottom line: AI can be a fascinating tool for testing strategies or spotting opportunities. But as with all things financial, a mix of caution, common sense, and a healthy respect for the unknown is still the best long-term plan. Nathan’s journey is one to watch—but it’s not one to blindly follow.
