How I Built an AI Trading Bot That Turned $25,000 Into $30,621

Name: How I Built an AI Trading Bot That Turned $25,000 Into $30,621
Uploaded: 2026-09-09T02:06:14.583Z
Duration: 51 s
Channel: Austin Starks
Description: A $25,000 public account traded by AI is up 22.49% in 126 days. The rules, the day-two failure that forced a risk-engine rebuild, and what the drawdown looks like.

Austin Starks deposited $25,000 into a public brokerage account in February 2026 and let an AI trade it. As of 8 September the account is worth $30,621, up 22.49% in the 126 days since the first trade, and ahead of the S&P by 8.33 points.

He wrote no code for it. The universe is the nineteen stocks already on his watchlist, chosen on conviction rather than screened. The strategy was described in plain English and refined by several AI models: buy at-the-money calls 365 to 730 days out on names trading above their 100-day average, ranked by 63-day rate of change, sized at 6% of the account each, gated on VIX under 35.

The part worth watching is day two. The bot bought a position he could not exit, because he had written an entry rule with no matching exit. He rebuilt the risk engine, and the book now carries nineteen per-company sell rules alongside a 300% take profit and a 180-day time stop.

Every figure comes from the live portfolio object, not from a backtest. The account has been down 12.21% at its worst and sits 5.61% below its August peak.

Austin Starks

Comments

Transcript

0:00Here's how I built the trading bot without writing a single line of code.

0:03I gave an AI twenty five thousand dollars of my own money. Today that account is worth thirty thousand six hundred and twenty one dollars.

0:11First I took the nineteen stocks on my watchlist. I have conviction in each of them for one reason or another, so that became my trading universe.

0:18Then I described the strategy in plain English. It buys long dated call options on the names above their hundred day average, ranked by momentum.

0:26I played Fortnite and checked in every so often, while different AI models kept applying changes and refining the strategy.

0:32On day two it bought a stock it should not have, and I could not sell it. I had written an entry with no matching exit.

0:39So now every company has its own rule that sells when the stock falls below its trend, plus a take profit at three hundred percent.

0:47If you want to learn how to deploy your own AI bot, leave a comment below.

Join the conversation

Loading conversation…