Last Tuesday afternoon, someone I've been mentoring for two years forwarded me a Telegram link — MoneySimpler's free AI trading bot, pitched for stocks, crypto and forex in a single wrapper. He asked if he should plug it into his Exness account. I asked him one question back. Does the bot report the commission per trade, or does it swallow the cost inside the spread? He didn't know. That is the tell. Exness's pro-account EUR/USD spread averages 0.1 pips. The standard-account spread averages 1.0. Every "free bot" pitch you will read this quarter lives inside that nine-tenths-of-a-pip gap. Let's open it up.

What Does "Free" Actually Mean When a Vendor Bundles an AI Trading Bot?

Free means the subscription line on your credit card is zero. It does not mean the round-trip cost of a trade is zero. Somewhere in the chain — bot vendor, broker introducer agreement, spread markup, or execution slippage — a revenue line has to exist, because nobody runs GPU inference for stocks, crypto and forex without a monetisation ledger somewhere.

There are three canonical shapes this takes. The first is an introducing-broker rebate: the bot vendor gets paid a cut of your spread by the broker for every lot you trade through them. The second is a widened spread on the specific instrument the bot prefers to trade. The third, the honest one, is a paid tier that unlocks features while the "free" tier is a lead magnet. When you cannot identify which of the three funds the vendor, assume it is the first two. That was the answer I gave my mentee. He thought "free" meant Exness's own numbers would apply. It usually does not.

Which Broker Model Is Cheaper for Bot-Driven Volume — Zero-Commission or Raw-Spread-Plus-Commission?

For any bot that trades more than about ten lots a month, raw-spread-plus-commission is cheaper. It is not close. The zero-commission model — XM's and Exness's marketing headline — hides its cost inside a wider spread. The raw-spread model, offered by Pepperstone and IC Markets on their standard commission-bearing accounts, quotes you interbank spread and adds a visible commission per lot. That transparency is what makes the math work at volume.

I want you to hold this in your head. At retail size — say one trade a week, 0.1 lot — the zero-commission model is almost always cheaper because commission is fixed per lot and spread is variable. At bot size — dozens or hundreds of lots a week — the arithmetic flips, because the spread markup on zero-commission accounts scales linearly with volume while the raw spread on commission accounts stays tight. Every bot vendor pitching "free AI" to a retail audience is banking on you not doing the crossover calculation. So we're going to do it.

Can an AI Trading Bot Legally Execute Trades Under FCA, ASIC and CySEC Rules?

Yes, but the responsibility for the trade sits with you, not with the bot. Under FCA CONC and COBS rules, ASIC's design-and-distribution obligations, and CySEC's transposition of MiFID II suitability language, the account holder is the decision-maker even when an algorithm places the order. A bot is a tool, legally indistinguishable from a spreadsheet macro. If it loses money, you cannot claim mis-selling against the broker because the bot vendor told you it would work.

There is a narrower question — whether the bot vendor itself needs to be authorised. If MoneySimpler markets the bot as an investment recommendation or manages your account on your behalf, that would fall under regulated investment advice or portfolio management, both of which require authorisation. If the bot is presented as software you install and control, it typically does not. The distinction is not cosmetic. Read the terms of service. If the vendor takes any discretion over your funds — even implicitly through auto-execution — the vendor is operating in a regulated activity space. If it does not have authorisation from at least one of FCA, ASIC or CySEC, that is a red flag independent of whether the strategy works.

How Do You Reproduce the Commission Math the Bot's Marketing Doesn't Show You?

Here is the calculation, in prose, that every marketing page skips.

Assume a bot places 50 round-trip EUR/USD trades a month, at 1 lot each. One standard lot is 100,000 units of the base currency, so 1 pip of EUR/USD movement is worth $10 per lot.

On Exness's standard zero-commission account, the average EUR/USD spread is 1.0 pip. That is $10 of cost per round-trip, per lot. Over 50 round-trips, the monthly spread cost is $10 × 50 = $500. Commission is zero. Total monthly cost: $500.

On Exness's pro account, the average EUR/USD spread is 0.1 pip. That is $1 of cost per round-trip, per lot. Over 50 round-trips, the monthly spread cost is $1 × 50 = $50. Commission is still zero on Exness's pro tier per the grounding data. Total monthly cost: $50. That is a $450 gap on the same broker between two account tiers a retail user rarely gets pointed toward.

On HF Markets' standard account, EUR/USD averages 1.2 pips — $12 per round-trip per lot, $600 monthly, no commission line. On HF's pro account, spread is 0.0 pips — you would still owe HF's standard commission structure per lot (not in our grounding, but historically $6-$7 per lot round-trip at commission brokers, which is what you should ask HF's disclosure page to confirm). Assume $6 per round-trip: $6 × 50 = $300 monthly. Still cheaper than HF's own standard account by half.

The bot's marketing shows you none of this. It shows you the win rate.

Does It Matter Whether the Bot Routes Through XM, Exness, Pepperstone or IC Markets?

Enormously. And the direction of the effect flips based on your trade size.

Exness's grounding shows a 1.0-pip standard EUR/USD spread and 0.1-pip pro spread, with a $1 minimum deposit and instant withdrawals. FBS shows a 0.7-pip standard, 0.0-pip pro, 1:3000 leverage, $1 minimum. AvaTrade sits at 0.9 pips on both standard and pro, with a $100 minimum deposit and conservative 1:400 leverage. FXTM shows 1.5 pips standard, 0.1 pips pro, with $10 minimum and Indian rupee account support. These are five very different cost surfaces for the same instrument.

If MoneySimpler's default routing is a broker whose zero-commission spread is 1.5 pips wide, and the same trade could have been placed at 0.1 pips on a pro-tier account with a small commission, you are handing 1.4 pips of edge per trade to the intermediary. On a bot doing 200 lots a month, that is $2,800 of gifted P&L. The question you must ask before you connect any bot: which broker does it default to, why that broker, and what is the vendor's compensation from that broker per lot you route? Vendors that will not answer question three are answering it by refusing to answer it.

What Happens to Your P&L When Spread Markup Replaces a Visible Commission?

You lose the ability to audit your own trading. That is the deeper cost, and it is worse than the dollar cost.

When commission is a separate line, you can pull a monthly statement, sum the commission column, and know exactly what you paid the broker. When cost is embedded in the spread, you cannot. Your entry price and exit price already have the vendor's cut baked in, and there is no invoice to reconcile against. If you want to know your true trading edge — pre-cost gross alpha — you have to reverse-engineer it from the average spread the broker publishes, which is itself a moving average the broker chooses. Two brokers publishing "1.0 pip average EUR/USD spread" may be applying that markup at very different points in the daily session, giving you very different real costs at your bot's actual execution times.

A visible commission is not a nuisance. It is the receipt that lets you catch the moment your edge disappears. When the marketing page says "zero commission" as a positive, translate it in your head to "no receipt for this cost." Then decide whether you want to trade blind.

Why Do Two Broker Cost Disclosures on the Same Question Contradict Each Other?

Look at Exness's own numbers in the grounding data. Standard account: 1.0-pip average EUR/USD spread, zero commission. Pro account: 0.1-pip average EUR/USD spread, zero commission. Both are the same broker's own disclosures. Both are operative. Both are true. They contradict each other only if you assume they should be consistent — and the reason they are not consistent is exactly what a bot vendor is trading against.

The standard account subsidises the pro account's tight spread through the wider markup on retail flow. The retail user paying 1.0-pip spread is, in an accounting sense, cross-subsidising the professional user paying 0.1. Both disclosures are honest at the tier level; both are misleading if a retail user thinks the standard account's headline cost is representative of the broker's cost base. The bot's marketing then compounds the mislead by quoting the pro-tier spread as achievable performance while auto-opening standard-tier accounts by default.

The primary-document cross-reference here is worth spelling out. Exness's public spread table is one document. Exness's account-tier terms and conditions is another. Read together, they describe two different products for two different customers. Read separately — which is how bot marketing presents them — they describe a mirage. Any bot vendor that quotes "0.1-pip spreads" in its pitch without specifying pro-tier eligibility (minimum balance, prior trading history, upgrade approval) is showing you the second document without the first. That is not a technicality. It is the whole difference between the number on the pitch deck and the number on your statement.

What Would Make Us Trust a Free AI Trading Bot Enough to Fund It?

Four conditions, none negotiable.

First, published broker routing. The vendor names which brokers it routes to, and lets you choose. Not a default that is hard to change, not "premium partners" — actual disclosure with named alternatives. If the bot only works with one broker, ask why, and assume the reason is an exclusive rebate arrangement until proved otherwise.

Second, the vendor publishes its own compensation. If MoneySimpler earns $X per lot routed through Broker Y, that number appears in the terms of service. It does not have to be zero. It has to be known. Regulated introducing brokers under FCA and CySEC frameworks already have to disclose this to authorised counterparties; a genuinely trustworthy free-tier vendor voluntarily extends the same disclosure to retail users.

Third, backtests on out-of-sample data with an execution model that assumes retail spread, not pro spread. If the marketing shows a Sharpe ratio, the fine print should show the same Sharpe ratio calculated at 1.0-pip cost per round-trip, not 0.1-pip. Almost none of them do.

Fourth, authorisation status on the vendor entity itself. Not on the broker it routes through — on MoneySimpler. If the software's marketing implies it will make discretionary decisions on your behalf, the entity making those decisions should be authorised somewhere.

We would reverse our skepticism if MoneySimpler published, in the same document, the broker routing map, the per-lot compensation received from each broker on that map, backtest performance calculated at retail-tier spread costs, and its own regulatory authorisation status. Until those four things exist on one page, the argument holds. Free is what the credit card statement says. It is not what your account statement will show three months in.

FAQ

How much does a "free" AI trading bot actually cost per month at realistic bot volume?

At 50 round-trip EUR/USD lots per month on a standard zero-commission account with 1.0-pip average spread, you are paying roughly $500 monthly in embedded spread cost — $10 per pip per lot, times 50 lots. On a pro-tier account with 0.1-pip spread, that same volume costs about $50. The "free" label refers to the vendor's subscription, not to your execution cost, and the gap between tiers is where the real bill lives.

Yes on the account-holder side. FCA, ASIC and CySEC treat the account holder as legally responsible for trades placed by any tool, algorithm or advisor they authorise. The narrower question is whether the vendor itself is offering regulated investment advice or discretionary portfolio management. If the bot makes trading decisions on your behalf, the vendor is likely operating in a regulated activity and should hold authorisation from at least one tier-1 regulator.

Which broker in this list is cheapest for a bot doing dozens of lots a week?

Based on the grounding data, Exness's pro tier at 0.1-pip average EUR/USD spread is the tightest cost surface for zero-commission execution, and FBS's pro tier at 0.0-pip spread is nominally tighter but should be verified against its commission schedule which is not in the grounding. AvaTrade at 0.9 pips on both tiers is the most expensive at volume. Retail volume flips the ranking; bot volume rewards raw-spread models.

What is the difference between XM zero-commission and Pepperstone standard for a high-volume bot?

XM's zero-commission model bakes the broker's revenue into the spread. Pepperstone's standard commission-bearing account quotes raw interbank spread and charges a separate per-lot commission. At high volume, the raw-spread-plus-commission model is almost always cheaper because interbank spreads compress with volume while zero-commission markups do not. The trade-off is transparency: with commission you can audit your cost; with markup you cannot.

How do I check whether a bot vendor is being paid by my broker?

Read the terms of service and the vendor's disclosures section. Introducing-broker relationships must be disclosed to counterparties under FCA and CySEC rules. If the vendor does not name its broker partnerships, ask them directly which brokers pay them a rebate and how much per lot. Refusal to answer, or vague "premium partners" language, is a signal that per-lot compensation is the vendor's actual revenue model.

Can I use one bot across stocks, crypto and forex on the same broker account?

Only if the broker itself supports all three asset classes on the account you have opened. In the grounding data, AvaTrade explicitly supports options and multi-asset via AvaOptions; Exness, FBS, FXTM and HF Markets are primarily forex and CFD brokers whose stock and crypto CFD coverage varies by jurisdiction and account tier. Check the specific instrument list before assuming the bot's "one wrapper" pitch matches your broker's actual instrument access.

Are Islamic (swap-free) accounts available for bot trading?

Yes. Every broker in the grounding data — AvaTrade, Exness, FBS, FXTM and HF Markets — offers an Islamic account option that removes overnight swap charges. Bots that hold positions overnight benefit materially from swap-free accounts because algorithmic strategies often carry positions through rollover. Confirm with the broker that Islamic-account status does not restrict access to the pro-tier spread structure, since some brokers scope swap-free to standard accounts only.