How Much Can an AI Trading Bot Make? Real Trade Examples After Fees

How much money can an AI trading bot earn?

The useful answer is a calculation, not a universal daily or monthly number. An agent may identify an opportunity and execute it automatically, but the economic result still comes from buying and selling at different prices, or from the reverse price move in a short position. Automation does not establish how large that result will be.

Start with the opening position value: the quantity traded multiplied by its entry price. A 1% return on a 100 USDT position is 1 USDT; on a 1,000 USDT position it is 10 USDT. Neither number tells you the return on the whole account without knowing the account equity and other positions.

We use USDT because that is the quote currency of the records. USDT amounts are not a promise of equivalent withdrawable dollars. The calculations exclude taxes, conversion and withdrawal charges.

Three documented trades, with their calculation basis

The source is TensorParadigm’s Telegram export reviewed on 10 October 2026. TensorParadigm confirmed that the posted prices were average exchange fills and that commission was 0.12% on entry and 0.12% on exit. The stated 0.03% slippage was already included in those prices. This is analysis of our published records, not an independent exchange audit.

These are the same three selected examples explained in our live-trade case study. That article provides the selection method and source context. The records do not supply traded quantities or a complete account-equity series, so the table below calculates returns on opening position value rather than claiming account growth.

QNT’s closing post displayed +9.88% with a 10× label; EDEN displayed +0.61% with a 3× label. Recalculating from the prices and the stated fees puts them on the common position-value basis below. NEAR’s closing post had no multiplier label. Funding cash flows were not supplied and are excluded throughout the historical examples.

What those returns mean for 100, 1,000 or 10,000 USDT

Each column independently replays one price pair at a hypothetical opening position value. These are arithmetic illustrations, not the amounts actually traded, recommended deposits or expected earnings. The 10,000 USDT column assumes the same fills and percentage fees as the smaller positions; larger orders may achieve different execution prices.

The table is after the stated trading commissions and before funding and other costs. Do not add these percentages to report a portfolio return. The positions have different entry times, and the account allocation and any overlapping exposure are unknown.

Worked example: the QNT trade on a 1,000 USDT position

The QNT entry was 253.72 USDT on 6 October 2026 at 07:17 UTC, and the exit was 256.84 USDT at 09:08 UTC. A hypothetical 1,000 USDT opening position corresponds to approximately 3.941353 units. Its closing value is 1,012.297020 USDT, giving 12.297020 USDT of price-move profit.

Entry commission is 1,000 × 0.0012 = 1.200000 USDT. Exit commission is 1,012.297020 × 0.0012 = 1.214756 USDT. The result after those commissions is 12.297020 − 1.200000 − 1.214756 = 9.882264 USDT, rounded to 9.88 USDT. The return on opening position value is 0.988226%, rounded to 0.9882%.

The exit commission uses the closing transaction value, which is slightly different from the entry value. Subtracting exactly 0.24 percentage points from the gross return would be a shortcut; calculating both transactions gives the more precise result. Exchange documentation describes commission as transaction notional multiplied by the applicable fee rate. [2]

AI trading bot earnings calculator

Load a documented price pair, choose an opening position value, and inspect the calculation. Enter your own fees when they differ from our examples. Optional fields let you include net funding paid or received, other costs, leverage and an account-equity reference.

The calculator models one constant-quantity long or short trade in a linear USDT-settled product, closed at the entered price. It does not model partial exits, position additions, minimum order sizes, liquidation, inverse contracts or a live order book. The historical presets contain prices, not an assumed future return. Changing a preset’s prices or direction creates a custom scenario.

From a trading result to spendable earnings

A trade can close at a favorable price while leaving less cash after costs. Use the actual commission on every fill. Subtract net funding paid and add funding received. Funding is a separate cash flow for perpetual positions held across the relevant settlement, rather than a second trading commission. [2] [4]

For the historical presets, leaving funding blank means the result is before funding; it does not assert that no funding payment occurred. If you enter 0, you are explicitly assuming zero funding for your scenario. The other-cost field can include a known amount allocated to that trade, such as a software or borrowing cost. Do not enter costs already included elsewhere.

Slippage needs the same discipline. When entry and exit are actual average fills, the execution difference is already in the prices. If you work from ideal signal prices instead, first adjust them to your assumed fills. Our examples do not deduct the owner-confirmed 0.03% again.

Why a 9.88% trading result is not automatically 9.88% account growth

For a simplified 1,000 USDT QNT position at 10× leverage, initial margin is approximately 100 USDT. The same 9.882264 USDT after commissions equals 9.882264% of that initial margin, while it remains 0.988226% of opening position value. This explains how the post’s rounded +9.88% can coexist with the smaller position-value return. [3]

If account equity before that trade were hypothetically 2,000 USDT, this one result would contribute about 0.4941% of that reference equity, before funding and other costs. That is a trade’s contribution, not a reconstructed account return. Other positions, open profit or loss, deposits and withdrawals still matter.

Holding position value fixed and changing leverage changes the illustrative margin denominator, not the price-move profit. Holding margin fixed and increasing leverage enlarges exposure, fees and possible losses. Our calculator keeps the entered position value fixed; its margin figure is not an exchange liquidation or cross-margin calculation.

A losing-price scenario makes the same arithmetic clearer

As an explicitly hypothetical example, keep QNT’s entry at 253.72 but set the exit 1% lower, at 251.1828. A 1,000 USDT opening position would lose 10 USDT from the price move and pay 2.388 USDT in total commissions at 0.12% per side. The result is −12.388 USDT, or −1.2388% of opening position value, before funding and other costs.

This is not a reported TensorParadigm trade. It is a sensitivity example available in the calculator. It shows why an earnings tool should let you test unfavorable exits as easily as favorable ones, and why a margin percentage magnifies a loss as well as a gain.

Can an AI trading bot make a reliable monthly income?

A few profitable trades cannot establish a recurring monthly amount. Monthly earnings depend on every position, its actual size, realized gains and losses, funding, trading costs and any fixed operating charges. Mark open positions consistently when measuring account performance; closed-trade profit alone can miss losses still held in the account.

Do not multiply the QNT example by 30 days, assume that each day offers the same setup, or compound the displayed Telegram percentages as if each trade used the full account. None of those assumptions is established by the supplied records.

A useful monthly review reconciles the complete exchange ledger and account equity, separates trading results from deposits and withdrawals, and compares outcomes across different market conditions. Positive months and losing months both belong in that review. Our passive-income guide explains the ongoing work behind automation.

Follow the agents before turning a percentage into an income expectation

TensorParadigm shares agent entries and exits free of charge in the public Telegram channel. You can observe the sequence of trades as posts appear, inspect the linked entry and exit records, and compare them with the methodology in our case study.

The three examples here demonstrate how to read earnings figures. They are not a complete ranking of our strategies or a guaranteed outcome for followers. Someone copying a post may trade a different size, incur different fees, enter later or receive a different fill.

The practical question is whether a system produces a useful result after costs at an exposure you can understand. Keep the prices, quantity, costs, leverage basis and account reference together. That turns “Can AI make money?” into a calculation you can check.