Best or Worst IdeaI Gave an AI $1,000 to Trade Stocks for 90 Days
Real money. No rules. Maximum gains. Let's see what happens.
The Question
Can an AI beat the market?
Not in a backtest. Not in a simulation. With real money, in a real brokerage account, making real trades.
I decided to find out.
The Setup
I gave Claude — Anthropic's AI — full control of $1,000 in my Robinhood account for 90 days.
Claude acts as my stock analyst. It scans the markets, analyzes catalysts, identifies high-conviction plays, and tells me exactly what to buy and sell. I execute within minutes.
This isn't a chatbot giving generic advice. This is an AI with a mandate: generate maximum gains by any legal means.
The Parameters
| Starting Capital | $1,000 |
|---|---|
| Timeframe | 90 days (Jan 30 – Apr 30, 2026) |
| Platform | Robinhood |
| Strategies | Long, short, options — anything legal |
| Rules | None. Maximum aggression. |
| Goal | Maximum gains. 2x, 5x, 10x. No ceiling. |
Why This Experiment?
Everyone talks about AI disrupting finance. Hedge funds use it. Quant firms use it. But what happens when a regular person hands an AI a small account and says "make me money"?
I wanted to test:
- Can AI identify high-probability trades in real-time?
- Can it react to catalysts — earnings, FDA decisions, short squeezes?
- Can it manage risk while still swinging for big gains?
- Can it beat the S&P 500 over 90 days?
This blog is the live documentation of that experiment.
The Rules of Engagement
Claude's Role (The Analyst)
- Scan markets daily for highest-upside opportunities
- Call every trade — what to buy, when, at what price
- Manage the portfolio — when to hold, trim, or rotate
- Use any legal strategy — stocks, options, leverage
- Maximize every dollar over 90 days
My Role (The Executor)
- Execute trades within minutes of Claude's call
- Report fills and end-of-day updates
- Check in every weekday at 7:10 AM Mountain
- No second-guessing — if Claude says move, I move
The Strategy
This isn't buy-and-hold. This is aggressive capital rotation.
- Concentrated bets — One or two high-conviction positions at a time
- Catalyst-driven — FDA decisions, earnings, short squeezes
- Fast rotations — Take profits quickly, redeploy into the next opportunity
- Options for leverage — When the setup warrants amplified exposure
- No sitting on cash — Every dollar should always be working
The goal is to compound gains through multiple winning trades, not wait for one big score.
What's at Stake
This is real money. I can lose it all. But the upside?
| Result | Portfolio Value | Return |
|---|---|---|
| Double | $2,000 | +100% |
| Triple | $3,000 | +200% |
| 5x | $5,000 | +400% |
| 10x | $10,000 | +900% |
The S&P 500 averages ~10% per year. We're trying to beat that in 90 days.
Follow Along
I'll be posting daily updates every weekday:
- What trades Claude called
- Why it picked them
- How the portfolio performed
- Running scoreboard vs. the starting $1,000
Bookmark this page. Follow me on social. Let's see if AI can really make money in the stock market.
Let's Go
Day 1 starts now.
$1,000. 90 days. No rules. Let's make history.
Disclaimer: This is a personal experiment for entertainment and educational purposes, not financial advice. Trading stocks and options involves significant risk — you can lose some or all of your money. I am not a licensed financial advisor. Do your own research before making any investment decisions. Past performance does not guarantee future results. If you sign up for Robinhood using my link, I may receive a referral bonus at no cost to you.