OpenAI official logo (public domain, Wikimedia Commons) — CryptoBriefing brand treatment
Research shows AI chatbots crowd into the same handful of tech and chip names, steered more by media buzz than by fundamentals
Ask an AI chatbot for stock picks and you may get a very short list. A new study finds that roughly 50% of the investment recommendations from ChatGPT and Gemini are concentrated in just 10 stocks.
The finding matters because more retail investors are treating large language models like a free financial advisor. If those advisors all read from the same script, the people following them could end up holding the same crowded bets without realizing it.
What the research found #
The headline result comes from a study released in November 2026. It looked at the stock recommendations generated by ChatGPT and Gemini and found that half of them clustered around a group of 10 names.
A separate working paper from the National Bureau of Economic Research (NBER) points in the same direction. That paper examined models including ChatGPT 5.0/5.2 and Gemini 2.5 Flash and analyzed the portfolios they built across 2025 and 2026.
The pattern was consistent. The models leaned heavily toward large-cap technology and semiconductor stocks, with ChatGPT showing a particularly strong tilt.
NVIDIA was the standout. ChatGPT often assigned the chipmaker between 18% and 20% of a portfolio, and concentration rose further in actively managed portfolios.
The sector math is even more lopsided. According to the NBER analysis, LLM-built portfolios held an average of 41% in semiconductor stocks. The S&P 500, by comparison, sits at roughly 20%.
AI, tech, and the markets they move—in one daily briefing.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
Gemini took concentration in a different direction. Its portfolios showed a median of only 4.5 stock holdings, with a preference for large firms and momentum stocks.
Following the headlines, not the balance sheet #
Stocks the models recommended received nearly 10 times the media coverage of average firms. The research found that AI recommendations tracked the volume of coverage more closely than they tracked underlying financial fundamentals.
The models also favored momentum stocks, meaning companies whose shares have already been rising.
The overlap across platforms adds another layer. Parallel experiments across different models produced consistent recommendations, with top holdings frequently shared between them.
As of November 2026, names that kept appearing near the top included NVIDIA (NVDA), Alphabet (GOOGL), Adobe (ADBE) and Broadcom (AVGO). Different chatbots, built by different companies, kept landing on many of the same tickers.
What this means for investors and the market #
Anyone using AI-generated picks may want to check how much of their portfolio ends up in a single sector or name. If the answer is 41% semiconductors and a fifth in NVIDIA, that is a deliberate bet, whether or not the investor meant to make it.
The research flags the risk of a disconnect between market performance and actual corporate health if models keep favoring hype over fundamentals.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.