How much money do you really need to start automated trading?

It’s the first question anyone asks before connecting a trading robot, and the one that gets the most inflated answers. The uncomfortable truth is that the minimum your broker accepts and the minimum it actually makes sense to trade with are two very different numbers. Here you’ll find concrete figures, real costs and one rule you should never skip.

Minimum capital for automated trading: technical vs sensible

Let’s start by clearing up a misunderstanding. When a broker advertises “accounts from €10,” it’s talking about the minimum deposit to open the account. That doesn’t mean you can trade sensibly with €10: it means you can get through the door.

There are actually three different numbers, and it pays not to mix them up:

• The broker’s minimum: what it requires to open an account. It’s a marketing number, not a trading one.
• The technical minimum: the capital the robot needs to open its smallest position (usually 0.01 lots), cover the margin the broker requires, and not get knocked out by a stop out on the first pullback.
• The sensible minimum: the capital at which that minimum position is a small risk for your account, and at which fixed costs don’t eat up any result you could possibly get.

A robot doesn’t work magic with account size: it executes rules. If your capital is so tight that each trade puts 10% of the account on the line, you don’t have a trading system—you have a countdown. The gap between the technical minimum and the sensible one is exactly what separates “being able to place orders” from being able to survive a losing streak—what traders call drawdown, and it will come, because it always does.

A worked example: gold and micro lots (0.01)

Let’s make it concrete with the instrument we know best: gold (XAU/USD).

• With most brokers, 1 lot of gold equals 100 ounces. A micro lot (0.01) is therefore 1 ounce.
• With gold trading at, say, around $3,300, that micro lot controls a notional value of about $3,300.
• For a retail client in Europe, maximum leverage on gold is usually 1:20. Opening that 0.01 lot therefore takes about $165 of margin. (Other regions may set different limits.)

But margin is just the ticket to get in. Every dollar gold moves means $1 of profit or loss per micro lot, and gold does move: sessions with ranges of $30, $50 or more are nothing unusual. A single 0.01 position can swing by tens of euros on a normal day.

Now do the math backwards, which is the right way to do it: if your system uses stops of, say, $15, each 0.01 trade risks about $15. For that to be 1–2% of your account—the range where prudent risk management usually sits—you need between €750 and €1,500. That’s the real arithmetic behind the “sensible minimum,” and the reason a €200 account that covers the margin on paper is still a fragile account in practice. For the step-by-step formula that turns your stop and your risk percentage into a lot size, see how to calculate lot size in gold.

The hidden costs of automated trading nobody tells you about

A robot’s returns don’t start at zero: they start in the red. Before it earns a single euro, the account has to pay for all of this:

• Spread: the difference between the buy and the sell price. On gold it’s usually around $0.20–0.40 per ounce, depending on the broker and the time of day. Every trade starts out losing that amount.
• Swap: the financing cost of holding positions overnight (usually charged triple on Wednesdays). For strategies that hold positions for several days, it adds up.
• Commissions: on ECN or raw accounts you pay a fixed commission per lot in exchange for a tighter spread. It’s neither better nor worse: it’s one more line to add up.
• VPS: if you want the robot trading 24/5 without depending on your computer, you’ll need a trading VPS: usually between €5 and €15 a month.
• The robot’s license or subscription: the cost of the software itself. Any serious provider—we do it at homeaureum.com—should publish it clearly before asking you for anything.

What matters isn’t each cost on its own but its relative weight: fixed costs are the same with €500 as with €20,000, so the smaller the account, the higher the bar the robot has to clear just to break even.

The golden rule: only capital you can afford to risk

Everything above is arithmetic. This is the important part: in leveraged trading you can lose all the capital you deposit, with or without a robot. No track record, no backtest and no metric prevents that.

That’s why the rule has no fine print: the only money that goes into a trading account is money whose total loss wouldn’t change your life. Not your emergency fund, not your rent money, not savings set aside for something specific. If an amount seems “small” to start with but losing it would really hurt, that amount doesn’t qualify either.

There’s also a practical reason, not just a prudent one: if the money hurts, you won’t let the system do its job. You’ll step in during a losing streak, switch the robot off at the worst possible moment and turn a temporary dip into a permanent loss. The capital you can handle emotionally is also the capital that’s statistically useful.

How to scale up your trading account step by step

The good news is that you don’t have to settle on the final amount on day one. The healthy way to size your account is in stages:

• Demo first. A few weeks watching the robot trade without real money teaches you its pace, how often it trades and what it does on the ugly days. For free.
• Small live account. A modest amount within your sensible minimum. The goal of this stage isn’t to make money: it’s to check that live execution (spread, slippage, swap) looks like what you expected.
• Months, not days. Before adding capital, let the system go through different market conditions, including a losing streak or two. It’s the only way to find out how your stomach handles the ugly part, not just your spreadsheet.
• Scale up in steps. If after several months the live numbers and your peace of mind hold up, increase in stages and review the risk per trade after each increase, not before.

And one nuance almost nobody mentions: scaling also works in reverse. If the trading keeps you up at night, reducing your capital isn’t failing: it’s fitting the system to its user. The goal isn’t to trade as big as possible today, but to still be able to trade a year from now.

How much money to start automated trading: summary table

With all of that on the table, here’s the map. Remember it’s a rough guide, not a recommendation:

Your exact figure depends on your broker, your system’s risk per trade and your personal situation. But if you take away just one idea from this guide, make it this one: don’t look for the minimum you can trade with; look for the minimum you can trade with well.