# Study Finds ChatGPT and Gemini Stock Picks Cluster in 10 Companies

By Enna Lee

Canonical URL: https://www.tokenpost.com/news/investing/53331
Published: 2026-11-10T21:16:42.000Z
Updated: 2026-11-10T21:16:42.000Z
Section: Investing

> A separate portfolio analysis found chatbot-built portfolios allocated an average 41% to semiconductor stocks, compared with roughly 20% for the S&P 500.

Stock recommendations from ChatGPT and Gemini clustered in 10 companies, while a separate portfolio analysis found chatbot-built portfolios allocated an average 41% to semiconductor stocks, raising concerns about similar concentrated bets among users.

A study released in November 2026 found that roughly half of the two chatbots’ stock recommendations fell within those 10 companies. A separate analysis examined portfolios built by large language models across 2025 and 2026. One result measured which companies appeared in recommendations; the other measured how much of a portfolio was assigned to a sector.

The portfolio analysis found that semiconductors made up an average 41% of chatbot-built portfolios, compared with roughly 20% of the S&P 500. That figure describes portfolio allocation, rather than the share of recommendations naming semiconductor companies.

ChatGPT portfolios often assigned NVIDIA (NVDA) 18% to 20% of their holdings. Concentration was higher in actively managed portfolios. The findings show how a single company can make up a substantial portion of a portfolio built from chatbot recommendations.

Gemini portfolios had a median of 4.5 stock holdings. They also favored large companies and momentum stocks, a term for stocks whose shares have already been rising. The two chatbots therefore showed concentration in different ways: ChatGPT portfolios often gave NVIDIA a large allocation, while Gemini portfolios had a small median number of holdings.

Recommendations across models frequently overlapped. NVIDIA, Alphabet (GOOGL), Adobe (ADBE) and Broadcom (AVGO) repeatedly appeared among top holdings as of November 2026. Companies included in the recommendations received nearly 10 times the media coverage of average firms, and the research found that the picks tracked media attention more closely than underlying financial fundamentals.

That overlap may matter for investors using chatbot-generated picks. If people receive similar recommendations and build portfolios around them, they could end up with exposure to the same companies and sectors. The research also flags a potential gap between market performance and companies’ fundamentals if models continue to favor media attention.