A 19-year-old engineering student from Japan has caught the internet's attention after claiming that he built an AI-powered trading bot in just 48 hours and generated $750,000 from an initial investment of only $68.

The student said they created the system using Claude Code, an AI coding tool developed by Anthropic, which they aim to use to automatically scan and analyze several cryptocurrency markets for trading signals.

The claim, which originated from a post on social media platform X, has sparked discussions about the growing role of artificial intelligence in cryptocurrency trading and automated financial systems. According to the post, the bot generated $6,732 on its very first night of operation, almost 99 times the amount used to start the experiment.

How did the 19-year-old student build an AI trading bot?

The student programmed the automated cryptocurrency trading system with the help of Claude Code, the viral post says. He decided to create a bot that would watch multiple exchanges and look for potential discrepancies in Bitcoin pricing, rather than having to keep an eye on the market manually.

The system reportedly tracks more than 50 markets at the same time and receives real-time Bitcoin price data from Binance.

It leverages this data to spot situations that give them the chance to acquire the same cryptocurrency at different prices in various markets.

An iPad was also reportedly used as a second monitor to keep track of the system's activity. The entire scene was reportedly completed in 48 hours, showcasing the potential of AI coding assistants for creating software projects in a relatively short time.

How does the AI trading bot work?

The reported trading system is based on a strategy known as cryptocurrency arbitrage.

It involves identifying price differences for the same digital asset across different trading platforms or markets.

For example, if Bitcoin is available at a lower price on one exchange and at a higher price on another, a trader may attempt to purchase it from the cheaper market and sell it in the more expensive one to earn a profit.

The bot reportedly automates this process through the following steps:

  • Market monitoring: It continuously tracks more than 50 markets to identify differences in cryptocurrency prices.
  • Live data collection: The system uses Binance's API to receive updated Bitcoin price information.
  • Opportunity identification: It analyses price movements and looks for potential arbitrage opportunities.
  • Automated trading: When the system identifies a suitable opportunity, it can execute trades automatically without requiring manual intervention.
  • Round-the-clock operation: The bot is designed to monitor market activity and look for opportunities throughout the day and night.

Since cryptocurrency prices can change within seconds, automated trading systems can help traders respond to market movements faster than manual trading methods.

How much money did the Claude Code trading bot make?

The most talked-about part of the student's claim is the reported profit of $750,000 from an initial amount of just $68.

The bot reportedly made $6,732 in its first night of trading. This is especially significant as it is about 99 times the initial amount.

The reported results have attracted attention from technology enthusiasts, cryptocurrency traders and people interested in AI-powered financial tools. The narrative has also sparked doubt over the ease of automated trading systems for individual coders, using AI coding assistants.

The reported $750,000 amount has not been independently substantiated, however. There are insufficient publicly verifiable trading records to determine if the amount is realized net profit or unrealized gains or another calculation.

Can Claude Code really build a profitable trading bot?

Claude Code is an AI code tool created by Anthropic to help users generate, comprehend, edit and handle software code. It can help developers develop applications, automate tasks, and create systems that can integrate with various APIs.

In this case, the tool was reportedly used to develop the trading logic and automate parts of the cryptocurrency trading process.

But coding a trading bot and developing a profitable trading system is another story. Automated bots can encounter multiple obstacles within the dynamic environment of live cryptocurrency markets.

Trading fees, sudden price fluctuations, limited liquidity, execution delays and differences between expected and actual prices can reduce potential earnings. A great deal that seems like it's a good deal on a screen might just not be there when the deal is done.

Thus, while AI-powered tools can accelerate software development and make it more accessible, they do not ensure successful trading or that financial risks are eliminated.