AI Trading: Success Rates & Profitability

Before the ChatGPT explosion, AI trading was just called “algorithmic,” and big banks owned it. Three AI-enhanced exchange traded funds (ETFs) hit the market as early as 2018. Now, the banking monopoly is fading as retail investors get their hands on AI software. But there’s a lot of uncertainty. Does AI trading work? How profitable is it? Should you spend time learning it?

All three AI-enhanced funds with 5-year histories outperformed the S&P 500 by roughly 6% from 2018 to 2023, suggesting AI is a strong long-term investor.

But the narrative is no fairy tale. Many question whether AI-enhanced trading is viable given fundamental risks it poses to the financial system. Others question whether AI can maintain results for retail investors once the market becomes saturated with machine minds.

Does it really work?

AI trading works moderately better than the S&P 500 over 3+ year periods and about the same as the S&P for shorter time frames. It helps to understand how trading works to wrap your head around this phenomenon.

Traders analyze securities using primary qualities and price trends over time. These techniques are known as as fundamental analysis and technical analysis, respectively. For example, a company’s profit is a quality and its stock price is a trend.

Artificial intelligence can easily learn how investors use these inputs to make decisions. There’s a correlation between information input and trader output that AI can replicate.

Better yet, AI can originate its own strategies for the future by analyzing historical relationships between inputs and successful buy/sell outputs. There’s no need to copy human traders. A rudimentary comparison is a regression curve in Excel’s line chart. AI can quickly formulate a curve to fit price trends like snakeskin.

AI trading works well enough to match the market and sometimes beat it, but it can also underperform when unforeseen events or imperfect correlations undermine its logic.

In other words, it performs a lot like humans. But that doesn’t mean we should throw the baby out with the bath water.

The reality is traders use AI to support decisions, not to outsource them. The role of AI is symbiotic with human traders, and it’s anything but clear-cut. This dynamic explains how copy trading platforms like eToro can exist, and why the trading profession is expected to grow 10% by 2031.

Human and machine traders walk hand-in-hand. One improves the other. AI trading works, and it’s worth your attention.

Success Rates

The best way to evaluate AI trading success is through exchange traded funds (ETFs).

AI-enhanced ETFs have outperformed the S&P 500 by an average of 5.91% over 5 years and 0.27% over 3-months. However, they underperform in the 1-month frame by -2.31% on average.

ETFs are vehicles in which retail investors like you and I can invest, meaning we can buy shares in ETFs like we can buy shares in Amazon. Fund managers pick primary securities like stocks, bonds, commodities, and currencies (FX), and the fund’s performance is a direct result of the primary securities’ performance.

Therefore, we can judge fund managers for their ability to pick winning securities.

AI-enhanced ETFs are those whose “picks” come from AI instead of humans. IBM was the first to launch an AI-enhanced ETF called AIEQ in 2017. Others followed suit, and performance data is insightful.

The table below shows return data against the S&P 500 Index for 11 AI-enhanced ETFs (IETC, IEDI, QTUM, ROBO, XT, KOMP, AIEQ, QRFT, AMOM, NVQ, DIP).

AI-Driven ETFFund Category1-Month Return
(vs S&P 500)
3-Month Return
(vs S&P 500)
5-Year Return
(vs S&P 500)
ROBOGlobal Small/Mid Stock-1.32%7.48%4.47%
XTMiscellaneous Sector-3.95%-0.13%7.83%
DIPLarge Value-0.93%
KOMPMid-Cap Growth-3.39%0.81%
AIEQLarge Growth-5.06%-2.24%5.44%
AMOMLarge Growth-1.58%-1.72%
NVQMid-Cap Value-1.95%-1.31%
QRFTLarge Growth-2.47%-1.76%
IEDIConsumer Cyclical-4.16%-3.50%
AI-Enhanced ETFs Returns vs S&P 500

Key Takeaway

No AI-driven ETF has underperformed the S&P 500 over 5 years. This suggests machines are good at long positions, where markets are more predictable and less susceptible to swings from unforeseen events.


It’s clear that AI performs well long term, but there’s the question of cost. There’s no point outperforming the S&P if your AI costs are higher than returns. As a retail investor, there are 2 ways you can incur costs from AI trading. Either you invest in a fund driven by AI, or you purchase AI software and bots to trade directly via a broker.

AI Trading is profitable when you invest in AI-enhanced ETFs over 5-year periods and when you can earn a about $1.42/trade using AI software on your own.

Profitability on ETFs

ETF profitability is equal to returns on invested capital minus fees. Fees are “wrapped” and listed as a net expense ratio. The good news is ETFs performance returns are always net of fees, so ROBO’s 5-year return of 4.47% is value after fees have been deducted.

This means $10,000 invested over five years results in $10,447 after fees, and the yearly growth rate is 0.88%. In reality, the fund actually grew by 1.83% per year but has 0.95% in fees.

Here’s a table displaying normalized results for ROBO, XT, and AIEQ.

In short, 5-year AI-enhanced ETFs have average annual returns of about 1.15%.

Profitability on Personal Portfolios

Actively managing an AI-enhanced portfolio isn’t as far-fetched as it sounds. You can even do it for free with the right broker and a minimum deposit. Software and bots are available for reasonable prices, and you can control the level of involvement to fit your strategy.

  • Brokers: $1 deposit
  • Screeners with alert assistance: $99.50/month
  • Textual Research: $49.66/month
  • Order router: $1.22/trade

A full-stack of AI trading software includes a broker, a stock screener with AI alerts, textual research AI, and an order router to automate trade execution between your screeners and the broker. In another article we looked at the 14 best AI software on the market. Average costs for each are shown above.

Here’s what your breakeven earnings would need to be on a monthly basis to cover the average costs of AI trading software:

Monthly Orders5006007008009001000
Break Even per Trade$1.52$1.47$1.43$1.41$1.39$1.37

In other words, you need to earn $1.52 per trade to break even with a volume of 500. With a volume of 700, the breakeven point goes down to $1.37. The higher your trade volume, the less you need to make. Assuming these trade volumes, the average is $1.42/trade.

Conclusion: should you learn it?

Now we know that AI trading works, has good long-term success rates, and can be profitable as an ETF or self-directed strategy. But it doesn’t seem to be all that better than human-driven strategies.

For the moment, it’s not. That doesn’t mean you should ignore it.

As AI improves over time, ROI will increase and so will the barriers to learn it. That alone is a strong enough reason get started. Add on the fact that AI generates profits already and you have very little to lose by investigating.

I’ve already started carving out 1% – 5% for AI trading.

Disclaimer: All Content on this site is information of a general nature and does not address the circumstances of any particular individual or entity. Nothing in the Site constitutes professional and/or financial advice, nor does any information on the Site constitute a comprehensive or complete statement of the matters discussed or the law relating thereto.

About the Author


Noah is the founder & Editor-in-Chief at AnalystAnswers. He is a transatlantic professional and entrepreneur with 5+ years of corporate finance and data analytics experience, as well as 3+ years in consumer financial products and business software. He started AnalystAnswers to provide aspiring professionals with accessible explanations of otherwise dense finance and data concepts. Noah believes everyone can benefit from an analytical mindset in growing digital world. When he's not busy at work, Noah likes to explore new European cities, exercise, and spend time with friends and family.


Scroll to Top