Teen Entrusts ChatGPT with $100 Investment, Sees Surprising Stock Market Gains in One Month

In June 2025, 17-year-old Nathan Smith from rural Oklahoma handed over control of a $100 investment portfolio to ChatGPT. Just four weeks later, the portfolio had surged by an impressive 23.8%, outperforming benchmarks like the Russell 2000 Index (+3.9%) and the biotech ETF XBI (+3.5%) during that same timeframe. Smith documented his experiment on Reddit, sparking interest across tech and finance circles, with media coverage from Decrypt, Futurism, and others.

However, Smith is careful to note: “This is not financial advice or a sales pitch; just a simple experiment I wanted to share.”

From $100 to a Compelling AI Investment Case

Smith’s project was inspired by the flood of ads promoting AI-driven investing triumphs. Wanting to test the hype, he instructed ChatGPT with strict guidelines: it could only purchase whole shares of U.S. micro-cap companies valued below $300 million. Smith manually executed the trades weekly based on ChatGPT’s recommendations, while Python scripts using Yahoo Finance monitored performance.

Add Cosmo Herald as a Preferred Source

Keeping everything transparent was essential. Nathan posted his method on GitHub and provided regular updates on Substack, enabling others to track or replicate the study. He emphasized that ChatGPT’s decisions were reviewed carefully—he intervened when the AI made contradictory suggestions or impossible trade requests.

ChatGPT’s Method: Focus on Micro-Caps, Stop-Loss Controls, and Risk Management

The experiment hinged on rigorous discipline. ChatGPT applied automatic stop-loss rules, quickly selling positions that dropped below certain levels. It also maintained a balanced allocation, spreading smaller bets instead of taking large, risky positions.

A notable success was CADL, a micro-cap stock that contributed nearly 50% of overall profits. The AI’s timely exit before a reversal demonstrated a cautious, almost intuitive approach to risk.

By the end of the month, the portfolio had increased by nearly $25 on the initial stake, marking a standout return compared to general U.S. market benchmarks.

Analyzing Risk-Adjusted Performance Metrics

Smith went beyond just tracking profits and evaluated professional risk indicators to assess the AI’s performance robustness.

Sharpe ratio: 0.94 — indicating solid risk-adjusted returns, just shy of the 1.0 benchmark considered very strong.
Sortino ratio: 2.00 — an excellent measure highlighting high gains paired with low downside volatility.

These metrics imply the AI approach wasn’t simply chasing risky bets but achieved returns with thoughtful risk management, at least within the limited timeframe.

Nevertheless, a single month isn’t sufficient to establish long-term dependability. Academic studies back this caution. Research from the University of Florida earlier this year showed that while AI-generated stock selections can outperform in simulation, their edge diminishes when applied to live markets or scaled investments.