Experiment 01 · AI Research Lab
A 12-month conviction-weighted $100,000 paper portfolio of nine AI-identified biotech acquisition targets, tracked live against the SPDR S&P Biotech ETF (XBI). The hypothesis: large language model screening can surface M&A candidates ahead of announcement.
Returns indexed to 0% at inception (May 1, 2026). Weighted daily closes vs. XBI ETF.
| Symbol / Company | Type | Shares | Current Price | Position Value | Daily | Total Return |
|---|
Entry prices are May 1, 2026 closing prices, fixed for the duration of the experiment. Portfolio inception capital: $100,000. Undeployed cash: $419.54. Prices via Finnhub; delayed up to 15 minutes during market hours.
Each week, the AI model may swap one holding for another from the broader biotech universe. Decisions are documented with rationale. The model can replace any single position per week; weights are redistributed to the replacement at current market prices.
| Date | Action | Ticker | Shares | Price | Rationale |
|---|
The portfolio was constructed using a structured AI-assisted screening process. The methodology is documented in full below.
Large language models can synthesize publicly available information (SEC filings, earnings transcripts, patent databases, and M&A comparables) to identify biotech acquisition targets with statistically meaningful alpha versus a passive biotech index over a 12-month window.
Starting universe: approximately 400 US-listed biotech and specialty pharma companies on NASDAQ and NYSE. Initial quantitative filters applied before AI scoring:
The AI model may execute one swap trade per week: selling one existing holding and replacing it with a new name from the broader biotech universe. The replacement inherits the prior position's weight. All trades are documented with explicit rationale. Weights are not rebalanced except through swap trades.
The portfolio is benchmarked against the SPDR S&P Biotech ETF (XBI), benchmarked at its May 1, 2026 closing price of $130.92.
The AI-generated thesis for each holding. Inception theses are not updated during the experiment; swap trade rationale is logged separately in the trade log above.
Important disclaimer. This portfolio is a research experiment conducted by The Fery Group to evaluate the capabilities of AI tools applied to financial analysis. No real capital is deployed. This page does not constitute investment advice, a recommendation to buy or sell any security, a solicitation, or an offer of any kind. The Fery Group LLC is not a registered investment adviser, broker-dealer, or financial planner. All data is sourced from public markets and may be delayed. Past results of any AI methodology do not predict future performance. Consult a licensed financial professional before making any investment decision.