AI already runs most of the volume on the exchanges. And that’s exactly why a beginner’s opening is wider here than in most other professions
If you're waiting for the moment artificial intelligence enters trading so you can get in ahead of it, you're about thirty years late. Here's why that barely affects you.
There's a sentence that gets repeated a lot and sounds reassuring: 'trading is one of the last professions AI hasn't reached yet.'
It isn't true. And you're better off hearing that from us than from the first person who decides to correct you in the comments.
Where artificial intelligence actually is right now
A few numbers that describe the situation without dramatising it:
- Algorithmic systems execute between 60% and 75% of volume in major equity markets. In the US the share is around 70%, and in currencies around 58%.
- High-frequency trading firms make up roughly 2% of trading firms but generate approximately 73% of equity volume.
- The Bank of England reports that automated systems now execute more than half of daily UK trading volume, up from about 35% in 2020.
This isn't a forecast for 2031. This is yesterday. Quant funds have used machine learning since the 1990s, and the first fully agentic hedge funds launched in 2026.
So the question 'when will AI enter trading' is the wrong one. The right question is: which part of trading did it enter — and what was left behind.
What the machines actually took
The answer is surprisingly specific. Algorithms won everywhere that speed wins.
Market making and spreads
Somebody has to stand on both sides of the market and earn a fraction of a cent across a million trades. That's a machine's job, and it was a machine's job long before you'd heard of trading.
Arbitrage
A price difference for the same asset in two places that lives for 40 milliseconds. A human physically cannot see it, let alone trade it.
Executing large orders
How to sell $400 million of stock without collapsing the price. That's a mathematical problem, and today an algorithm solves it.
Processing news
A machine reads an inflation report and reacts before you've opened the tab.
Notice what all four have in common. Not one of them was a job a beginner could have done. You lost a race you were never in.
Now the number almost nobody quotes
In May 2026, a survey of 131 asset managers found that only 6% let AI make the final investment call. 74% use it for back-office work.
Read that carefully, because it's the whole article in one sentence. The industry with the best models, the best data and the best engineers on the planet uses AI almost everywhere except for deciding what to buy and when.
Machines won the microsecond. The decision about the next three days is still made by a human — including inside the institutions that could afford otherwise.
The reason isn't sentimental. Markets aren't a closed system like chess. The rules shift, the data is noisy, and any model that starts working too well changes the very market it was trained on and stops working. That's a hard problem, and it isn't solved.
And here's the real difference from other professions
This is why your instinct was right, even if for a different reason.
When AI arrived in translation, accounting or copywriting, it competed for the same work you were paid to do. The client has a choice: you or the model. If the model is good enough and cheap enough, your seat is gone.
Trading doesn't work that way. The market is not an employer.
Nobody can fire you from the market. There's no client deciding they'd rather use an algorithm than you. The fact that a firm in London executes 50,000 trades a second does nothing to stop you opening a position, holding it four days and closing it. You aren't competing with them for one job. You're simply in the same market.
That's a rare structure. In few other fields can you work inside a machine-dominated industry without the machines taking your place.
The good consequence and the bad one
We've covered the good one. The bad one deserves the same clarity.
The absence of a threat is not the presence of an edge. Between 89% and 95% of people who trade actively lose money within their first year. That figure looked the same before the algorithms and after them. Machines aren't why people lose. People lose because most of them have no process.
So AI isn't taking your chance. It isn't handing you one either. Your chance depends on the same things it depended on thirty years ago.
Where room for you actually remains
Three zones where the machine advantage barely applies.
1. The longer horizon
High-frequency systems win in windows measured from milliseconds to minutes. A trade you hold for three days or three weeks sits outside the range of their advantage. There, reading context decides the outcome, not speed.
2. Small size
This sounds like a weakness and it's your most underrated advantage. A large fund can't take a €2,000 position — it isn't worth their costs. You can enter and exit instantly without moving price, and sit out for weeks without answering to anyone. Large capital has no such freedom.
3. The right not to trade
A market maker is obliged to quote. A fund is obliged to be invested. You owe nobody anything. Being able to stand aside when conditions aren't yours is a structural advantage institutions don't possess.
One warning about AI tools for traders
The market for AI tools aimed at individual traders is growing fast — roughly 12.8 million users and over 40% growth in a year.
Worth saying plainly: buying an AI tool is not an edge. An edge is by definition something others don't have. A scanner that two hundred thousand other people can also access is just another screen. It may save you time. It will not make you profitable.
What to do with this if you're starting out
- Stop competing on speed. Don't trade the first minute after a news release. There you're losing to an opponent six orders of magnitude faster than you.
- Pick a horizon the machines don't serve. A daily or four-hour chart is entirely sufficient to begin with.
- Use your small size deliberately. Trade rarely. Wait. That's a freedom funds pay millions for.
- Put the effort into process, not tools. A written rule for risk per trade, one setup, a trading journal. That's what separates the 10% from everyone else, and it isn't a technology problem.
Artificial intelligence isn't coming for trading. It has been there for decades and its territory is clearly marked. The territory left for you is clearly marked too — it just asks for discipline instead of speed. The good news is that discipline is the only item on this list that can genuinely be learned.
Risk warning: This content is educational and does not constitute investment advice or a personal recommendation. Trading financial instruments carries a high risk of capital loss and is not suitable for everyone. A significant proportion of retail traders lose money. Never trade with funds you cannot afford to lose. Market structure data as of October 2026.