AlphaPicks

AlphaPicks

Share this post

AlphaPicks
AlphaPicks
Why The Perfect AI Hedge Fund Doesn't Exist
Copy link
Facebook
Email
Notes
More

Why The Perfect AI Hedge Fund Doesn't Exist

Creating the ultimate trading machine hasn't been perfected, with most AI funds actually underperforming.

AlphaPicks's avatar
AlphaPicks
Oct 18, 2023
∙ Paid
11

Share this post

AlphaPicks
AlphaPicks
Why The Perfect AI Hedge Fund Doesn't Exist
Copy link
Facebook
Email
Notes
More
3
Share
Can AI-Powered Automation and Human-Centered Customer Service Co-Exist?

Artificial intelligence (AI) is definitely going to appear on the bingo sheet for top words of 2023. While the focus has been on Nvidia and other tech names pushing the frontiers in this capacity, investing and trading has been putting AI to use for several years.

Granted, it’s an ever evolving phrase that encompasses more and more. But on the face of it, it doesn’t seem like the finance world has quite mastered the art of AI and profit generation.

Evidence of this can be seen from the below returns over the past few years of AI hedge funds versus the main index:

How funds are using AI

There’s a huge spectrum of uses of AI in trading right now. For example, we wrote back in April about J.P Morgan releasing an AI model that poured over Fedspeak to try and assign probabilities related to monetary policy actions:

JP Morgan Has A New AI Trading Model...And It's Good

AlphaPicks
·
April 28, 2023
JP Morgan Has A New AI Trading Model...And It's Good

The Hawk-Dove Score analyses language from Fed officials and rates it as hawkish or dovish. It then can provide a trading signal in the lead up to the next Fed meeting, if there’s a mismatch in language versus expectations. Although unlikely to be made available to the public, we flag up other trading models that do work.

Read full story

Another angle of AI is in more traditional quant modelling. At it’s most basic form, you could be forgiven for saying that a quant trading strategy is a quasi-form of AI. These type of funds have been around for decades.

Further, AI can be used not to actively trade, but rather to cut out the pre-trade research. The ability for it to digest information via machine learning can eliminate hundreds of manual hours of research done by analysts instead.

Yet the more recent nuance is tweaking quant models to include deep-learning elements. While a quant model could say “buy a stock if earnings per share are greater than X”, deep learning could figure out that this only works in some areas of the market, or factor in broader market sentiment and expectations.

The levels to this can be noted in the below graphic.

Simplifying the Difference: Machine Learning vs Deep Learning - Singapore  Computer Society

For the rest of this article, we’ll discuss:

  • The 3 major reasons why the average AI hedge fund underperforms

  • The top funds that are worth watching for the future

Reasons for underperformance

We identify three main reasons for AI funds dragging their feet.

1, Noisy Markets

This post is for paid subscribers

Already a paid subscriber? Sign in
© 2025 AlphaPicks
Privacy ∙ Terms ∙ Collection notice
Start writingGet the app
Substack is the home for great culture

Share

Copy link
Facebook
Email
Notes
More