AI Stock Picker: ChatGPT’s $500 Investment Portfolio Defies Expectations

In a bold experiment pushing the boundaries of artificial intelligence and financial markets, a journalist tasked ChatGPT, specifically its advanced GPT-5 model, with an unusual challenge: invest $500 of real money with the sole aim of maximizing returns over a six-month period. The results were far from the conservative, predictable advice one might anticipate from a sophisticated AI. Instead, ChatGPT delivered a portfolio characterized by its cleverness and aggressive nature, eschewing safety for high-growth potential.

The AI Investment Experiment: A Departure from the Norm

The premise was simple yet intriguing: provide an AI with a modest sum and a clear objective – significant profit in a short timeframe. This endeavor was not the author's first foray into testing AI capabilities. A previous experiment in 2022 with OpenAI's GPT-3 model, while yielding positive results that outperformed the market, was described as rudimentary. GPT-3's strategy then involved a momentum-based approach, favoring stocks that had already shown upward trends, alongside established tech giants like Microsoft, Apple, and Amazon. While successful, the gains were modest relative to the timeframe, and the investment landscape of 2022 was generally more favorable, with broad market growth benefiting most individual stock picks.

Entering 2025, the investment climate is considerably more complex, marked by geopolitical uncertainties, tariff tensions, and the pervasive influence of artificial intelligence itself. Against this backdrop, the author approached GPT-5 with a specific prompt: "I will give you $500 to invest in the stock market. You may choose up to 5 stocks. Make your picks, explain why, and I will buy them and we will see how they do." The expectation was a cautious response, perhaps a deflection towards professional financial advisors or a recommendation for a low-cost index fund – typical of AI models that have become more risk-averse due to increased public usage and safety protocols.

ChatGPT's Surprising Portfolio: Aggression Over Safety

However, GPT-5 defied these expectations. After an eight-minute deliberation, it presented a "Diversified High-Growth Portfolio," a selection that was anything but generic or timid. The AI recommended an even distribution of the $500 across five distinct companies: Palantir (PLTR), AppLovin (APP), Agios Pharmaceuticals (AGIO), Hut 8 Corp. (HUT), and MicroStrategy Inc. (MSTR). This selection immediately stood out due to its unconventional nature; several companies were unfamiliar to the author, and even those recognized, like Palantir, were not typically considered by the author for investment.

What made this selection particularly compelling was GPT-5's transparent articulation of its investment thesis. Unlike its predecessor, GPT-5 detailed its reasoning, citing research from 98 articles and websites. For Palantir, the AI highlighted its AI and data platform, noting its growing traction in both commercial and government sectors and the strong investor enthusiasm for its AI-driven growth. AppLovin was favored for similar reasons, specifically its "proprietary AI engine," with the AI also pointing to analyst notes suggesting room for growth despite recent gains.

Agios Pharmaceuticals presented a different kind of bet. GPT-5 identified the company as awaiting a crucial Food and Drug Administration (FDA) decision on expanding its drug Pyrukynd's approval for treating thalassemia. The AI recognized the significant unmet need and the potential for a substantial rally if the FDA decision was positive, framing it as a high-risk, high-reward play contingent on regulatory approval.

The inclusion of Hut 8 Corp. and MicroStrategy Inc. signaled a deliberate move into the cryptocurrency space. GPT-5 noted MicroStrategy's substantial holdings in Bitcoin, positioning it as a "highly leveraged Bitcoin play." Hut 8 was recognized for its evolution from a crypto miner to an energy-infrastructure platform supporting Bitcoin mining and AI/HPC data centers. The AI explicitly concluded that the portfolio