Is AI Trading Profitable? Our Live Results, Losses Included

Is AI trading profitable? AI trading can be profitable, but only when a model proves it with live trades, losses included, and most individual traders who trade often still lose money. On November 10, 2026, two of TruthSayer AI's three live-traded models were ahead of the S&P 500 since inception, and one was behind.

AI trading means using statistical or machine-learning models, trained on past market data, to decide when to buy or sell. "Profitable" means the trades made money after costs, over enough trades and enough time to rule out luck.

This guide is for self-directed investors deciding whether an AI trading app, bot or signal service is worth it. You will see our live results with the losing trades counted, what the research says about who makes money, and a six-step test to check any AI's profits before you pay.

What is AI trading, and how does it make money?

AI trading makes money the same way any trading does: it buys for less than it sells, or sells short for more than it buys back, more often or by more than it loses. The AI part is the model that picks the trades. It learns patterns from past prices, volumes, filings and other market data.

Products sold as AI trading fall into four main types. The type decides who places the trade and where your money sits:

• Signal apps. The app sends a buy or sell alert, and you place the trade in your own brokerage. TruthSayer AI is a signal app.

• Auto-trading bots. Software places orders for you, in an account you connect or on a platform that holds your money.

• Stock scores and screeners. The app ranks stocks, and you decide what to do.

• AI chatbots. General tools such as ChatGPT that explain stocks but do not test or trade anything.

Is AI trading profitable in 2026? What the data says

AI trading is profitable for a minority of well-tested systems, and the clearest winners are firms, not individuals. The best 2026 data comes from India's market regulator, which tracks the profit and loss of every type of trader in equity derivatives.

In its August 20, 2026 study, the Securities and Exchange Board of India (SEBI) found that 87.7% of individual traders lost money in fiscal 2026 (the year to March 2026). Proprietary trading firms made the largest gross profits, and 99% of the profits of proprietary traders and foreign portfolio investors were made by "algo entities", firms that used at least one algorithmic order that year.

Who made money in India's equity derivatives market, FY26

So algorithmic trading clearly can make money. The open question for you is whether a product you can buy carries that edge, or just the label.

Does AI trading work for individual traders?

For individual traders, the evidence says trading more usually earns less. SEBI found that higher trading intensity went with higher loss rates, and that loss rates fell from 93% for traders with no stock holdings to 58% for those with portfolios above ₹10 crore.

The same pattern shows up in older, larger studies:

• Brazil. Of all individuals who started day trading Brazilian equity futures from 2013 to 2015 and kept going for more than 300 days, 97% lost money (Chague, De-Losso and Giovannetti, 2020).

• United States. Across 66,465 households at a US discount broker from 1991 to 1996, the households that traded most earned 11.4% a year while the market returned 17.9% (Barber and Odean, 2000).

"Our central message is that trading is hazardous to your wealth."

— Brad M. Barber and Terrance Odean, Journal of Finance (2000)

How profitable has TruthSayer's AI been in live trading?

TruthSayer AI's live results are mixed in the way real results usually are: two models were ahead of the S&P 500 and one was behind, as of November 10, 2026. Every alert our models generate is traded in a live funded account, and the product pages list every trade, wins and losses.

Is AI trading profitable? TruthSayer live ROI vs the S&P 500, Nov 10, 2026

TruthSayer model, liveS&P 500, same dates

The losing trades are part of the record. Here is what each live page showed on November 10, 2026:

Live model	ROI since inception	S&P 500, same dates	Difference	Win rate	Closed profitable	Closed loss	Average hold
Valuations AI©	78.09%	31.94%	+46.15 pts	72.1%	151	45	78 days
Dark Pools AI©	33.62%	29.44%	+4.18 pts	67.3%	206	93	46 days
Earnings AI©	31.85%	35.09%	−3.24 pts	72.8%	1,044	361	27 days

The three lists name 45 + 93 + 361 = 499 closed losing trades. Win rates, trade counts and holds are shown exactly as published on each page that day; the differences are simple subtraction (78.09 − 31.94 = 46.15).

How do you check the profit on a real AI trade?

You check an AI trade's profit by redoing the math from its entry and exit prices, then sizing it to your own account. Here is one real losing trade and one real winning trade from TruthSayer's live lists, read on November 10, 2026:

1. Find the losing trade. Tesla (TSLA), Dark Pools AI©, long. Entry $411.21 on July 10, 2026; exit $368.60 on November 5, 2026; held 118 days.
2. Check the math. $368.60 − $411.21 = −$42.61 per share. −$42.61 ÷ $411.21 = −10.36% raw price change. The list shows −10.97%, 0.61 points lower.
3. Size it. With 5% of a $10,000 account in the trade, that is $500. $500 × −10.97% = −$54.85, or −0.55% of the account.
4. Find a winning trade. Cloudflare (NET), Earnings AI©. Entry $241.77 on November 5, 2026; exit $269.25 on November 6, 2026. $27.48 ÷ $241.77 = 11.37% raw; the list shows +10.68%.
5. Size it. $500 × 10.68% = +$53.40.
6. Net the two. −$54.85 + $53.40 = −$1.45. One win of the same size only cancels one loss.

The published figures sit below the raw price changes on both trades. Ask any provider what its published returns include, and expect a clear answer. Our stock profit calculator runs the same math for any entry and exit.

Does a high win rate mean AI trading is profitable?

No. A high win rate does not make AI trading profitable; the size of the average win against the average loss does. This is the most common mistake in AI trading marketing, where win rates of 60% to 80% are presented as proof on their own.

Our own live data shows why. On November 10, 2026, Earnings AI© had the highest win rate of TruthSayer's three models (72.8%) and was the only one behind the S&P 500 (31.85% against 35.09%). Dark Pools AI© had the lowest win rate (67.3%) and was ahead (33.62% against 29.44%).

Earnings AI© still made money. Its page lists a beta of 0.30, which means it moved far less with the market than the index did, so trailing a strong market is not the same as losing.

How do you calculate whether a strategy makes money?

Use expectancy: (win rate × average win) − (loss rate × average loss). Illustrative example, not TruthSayer data:

• 70% wins averaging +4%, 30% losses averaging −8%: (0.7 × 4) − (0.3 × 8) = 2.8 − 2.4 = +0.4% per trade.

• Same 70% win rate, losses averaging −10%: 2.8 − 3.0 = −0.2% per trade. Same win rate, now a losing strategy.

Real traders fall into this trap. In SEBI's FY26 study, nearly 79% of traders who had both profitable and losing quarters made smaller average gains in their good quarters than they lost in their bad ones.

Why do AI trading backtests look better than live results?

AI trading backtests look better than live results because a model tuned on past data partly fits noise, and real markets add costs and competition. A backtest is a run of a strategy on past data; it shows what might have happened, not what did.

The best-known measure of this decay comes from finance professors R. David McLean and Jeffrey Pontiff, who tracked 97 published predictors of stock returns:

How much of a backtested edge survives? 97 published stock predictors

"Portfolio returns are 26% lower out-of-sample and 58% lower post-publication."

— R. David McLean and Jeffrey Pontiff, Journal of Finance (2016)

How does TruthSayer test a model before trusting it?

TruthSayer tests each model on data it never saw, against chance, and then removes the signals that fail. For Dark Pools AI©, retrained on 483 S&P 500 stocks with about two years of data, the research backtest (before trading costs and slippage) showed:

• Unseen data. 62.7% accuracy on the final 20% of the data, held back until final scoring, against a 50.9% baseline (65.5% in training).

• Shuffle test. 0 of 200 runs with scrambled outcomes matched the real model's 65.0%; they averaged 51.3%, with a best of 52.6%.

Dark Pools AI©, strictest filter (research backtest, before costs and slippage, outliers removed)	Trades	Win rate	Mean return per trade	Both tests passed?
Price had moved	165	72.1%	+0.92%	Yes
Price was flat	304	49.0%	+0.08%	No
Combined	328	54.0%	+0.25%	No

We require a t-test and a binomial test to both pass at p < 0.05. Only the "price had moved" group passed, so the flat-price signals are not part of the validated signal. Keeping them would have cut the mean return per trade from +0.92% to +0.25% in this research backtest.

Research and live figures measure different things: a research test and live trading. The 72.1% above is a research backtest win rate; the live Dark Pools AI© win rate was 67.3% on November 10, 2026.

What counts as a good result for an AI trading service?

A good AI trading result is a live, dated record that beats the S&P 500 over the same dates, shows its losses, and has been checked by someone independent. Use this decision guide:

Check	Good sign	Average	Warning sign
Track record	Every live trade, dated, losses included	Monthly summary only	Backtests or screenshots only
Benchmark	S&P 500 over the same dates	Benchmark over different dates	No benchmark
Win rate	Shown with average win and loss	Win rate alone	90%+ with no losses listed
Sample	Hundreds of closed trades over 2+ years	A few dozen trades	A handful of trades or one hot month
Verification	Independent, such as GIPS® verification	Self-reported full trade list	None
Promises	Past results, with risks stated	Vague "high accuracy"	Guaranteed returns

Is AI trading worth it compared with the alternatives?

AI trading is worth it only when it beats what you could do without it, after its cost. The honest default is an index fund, which tracks the market at low cost and needs almost no time.

Option	How profit is proven	Your time	Main risk	Best for
Index fund	Index data, public	Minutes a year	Market falls	Long-term savers who want the market return
Do-it-yourself trading	Your own records	Hours a week	Overtrading and bias	Experienced investors with a tested process
Stock newsletter	Varies; often picks without exits	An hour a week	Hidden losers	Readers who want ideas, not signals
AI signal app with a live record	Live trade list with losses	Minutes per alert	The model can stop working	Self-directed investors who will audit results
Auto-trading AI bot	Often backtests only	Setup, then monitoring	Fast losses; custody scams	Technical users who can test code

Each option has a real case. An index fund wins on cost and effort. A signal app with a live record wins on proof you can check. A bot wins on speed but is the hardest to verify.

Best for long-term, self-directed investors

AI signals with holds of weeks suit investors who check their account once or twice a day. TruthSayer's average live holds ran from 27 to 78 days on November 10, 2026.

Best for earnings traders

Earnings-driven signals suit traders who can act within hours. Earnings AI© lists a 1–4 hour entry window before a company reports.

Best for beginners

Beginners should start with an index fund and paper-track an AI service's live alerts for one to three months before risking money. Watch how losses feel, not just how wins look.

Who is AI trading not right for?

• Day traders chasing speed. The Brazil and SEBI data show high-frequency individual trading loses for most people.

• Anyone who needs this money within a year. Every model here has losing trades.

• Small accounts where fees eat the edge. See the cost table below.

What does AI trading cost, and how much time does it take?

AI trading costs a subscription plus trading costs and taxes, and the subscription matters most on small accounts. TruthSayer costs $71.88 a year ($5.99 a month billed annually) or $10.99 a month ($131.88 a year), after a 7-day free trial, as listed on November 10, 2026.

Account size	Annual plan ($71.88) as % of account	Monthly plan ($131.88/yr) as % of account
$5,000	1.44%	2.64%
$10,000	0.72%	1.32%
$25,000	0.29%	0.53%
$100,000	0.07%	0.13%

Each figure is the yearly fee divided by the account size; for example, $71.88 ÷ $10,000 = 0.72%. Your expected edge has to clear that bar every year.

Taxes count too. In the US, gains on stocks held one year or less, such as the 118-day Tesla trade above, are short-term capital gains taxed at ordinary income rates, according to the IRS.

Time depends on the entry window. On TruthSayer's pages, Valuations AI© allows 1–2 days to enter, Earnings AI© 1–4 hours and Dark Pools AI© 1–2 hours. After you enter, the work is tracking the position until the exit alert arrives.

What are the risks and common mistakes in AI trading?

The biggest risks in AI trading are trusting the wrong evidence and sizing trades too big. These five mistakes cause most of the damage:

1. Judging by win rate. A 70% win rate can still lose money if losses are bigger than wins.
2. Trusting backtests. Published stock predictors lost 58% of their return after publication (McLean and Pontiff, 2016).
3. Oversizing. At 5% per trade, the Tesla loss cost 0.55% of the account; at 50%, the same trade would have cost 5.49%.
4. Overtrading. The most active US households earned 11.4% a year against the market's 17.9% (Barber and Odean, 2000).
5. Ignoring costs and taxes. Fees and short-term tax rates come off every winning trade.

The TruthSayer Profit Proof Test: how do you check any AI's profits?

The TruthSayer Profit Proof Test is a six-step check you can run on any AI trading app, bot or signal service in an afternoon. It uses only what the provider publishes:

1. Live, not backtested. Find real-money trades with dates, prices and losses. Backtests alone do not count.
2. Same dates, same benchmark. Compare total return with the S&P 500 over the exact same period.
3. Expectancy, not win rate. Work out the average win, the average loss and how often each happens.
4. Your costs at your size. Divide the yearly fee by your account size, and add trading costs and taxes.
5. One loss, in dollars. Apply the worst published loss to your position size and decide if you can live with it.
6. Someone else checked. Look for independent verification, such as GIPS®, and note the period it covers.

How does TruthSayer AI apply this to its own signals?

TruthSayer AI applies the same standard to itself: live trades first, losses published, and independent verification of how performance is reported. Every alert from Valuations AI©, Earnings AI© and Dark Pools AI© is traded in a live funded account before you see it. TruthSayer does not hold your money or place trades for you.

TruthSayer AI claims compliance with the GIPS® standards, and its firm-wide policies have been independently verified for September 1, 2024 to June 30, 2026. The verifier's reports are on our GIPS page, and our guide to GIPS compliance explains what that covers.

Research decides what reaches you. Signals that fail our tests, like the flat-price Dark Pools AI© group in the research backtest above, are left out.

What should you do next?

Run the TruthSayer Profit Proof Test on any service you are considering, starting with its losing trades. Then decide with your own numbers:

• If you have under $10,000: start with an index fund and paper-track one AI service until the fee is under 1% of your account.

• If you hold for weeks: compare live records of signal apps that list every trade, such as the Valuations AI© and Dark Pools AI© pages.

• If you trade earnings: read the Earnings AI© trade list and check how it did against the S&P 500 over the same dates.

To go deeper, read whether AI trading is legit and how to spot a fake platform, why backtested and live-traded signals differ, and how to judge a signal service. Our guide to the best AI trading app compares apps on price and proof, and the investing glossary defines every term.

Frequently asked questions

Is AI trading profitable for beginners?

Rarely on its own. SEBI found that 87.7% of individual equity derivatives traders in India lost money in FY26, and the loss rate was 93% for traders with no stock holdings. Beginners do better starting with an index fund, then paper-tracking an AI service's live signals for one to three months.

How much money do you need for AI trading?

You need enough that fees stay small. TruthSayer's annual plan costs $71.88 a year as of November 10, 2026, which is 1.44% of a $5,000 account but 0.29% of a $25,000 account. With 5% position sizes, a $10,000 account puts $500 into each trade.

Can AI trading beat the S&P 500?

Some AI models have, over set periods, but not all. On November 10, 2026, TruthSayer's Valuations AI© showed 78.09% since inception against 31.94% for the S&P 500, while Earnings AI© showed 31.85% against 35.09%. Past results do not guarantee future results.

What should I do after I start following AI trading signals?

Log every trade you take: entry, exit, position size and fees. After 20 to 30 closed trades, compare your return with the provider's published trades and the S&P 500 over the same dates. If your results trail the published list by more than its fee, check your entry timing.

How often should I review an AI trading service?

Review it every quarter. Each time, check its live return against the S&P 500 over the same dates, how many closed losses it lists, and whether its independent verification has been extended. TruthSayer's current GIPS® verification covers September 1, 2024 to June 30, 2026.

Is AI trading worth it?

AI trading is worth it only if its live edge, after costs and taxes, beats an index fund at your account size. Check three numbers: live return against the S&P 500 over the same dates, the yearly fee as a share of your account, and how many losing trades the provider publishes.

Sources

• TruthSayer AI, Valuations AI© live results and trade list (accessed November 10, 2026)
• TruthSayer AI, Dark Pools AI© live results and trade list (accessed November 10, 2026)
• TruthSayer AI, Earnings AI© live results and trade list (accessed November 10, 2026)
• TruthSayer AI, pricing (accessed November 10, 2026)
• TruthSayer AI, GIPS® compliance and verification (verification period September 1, 2024 to June 30, 2026)
• TruthSayer AI, Dark Pool AI Signal Statistical Validation Report, S&P 500 retrain (internal research backtest, summary figures only)
• SEBI, press release PR No. 50/2026 on the profitability and trading behaviour of individual traders in equity derivatives (FY25–FY26) (August 20, 2026)
• SEBI, Study: Profitability of Individual Traders in the Equity Derivatives Segment (FY25–FY26) (August 20, 2026)
• McLean and Pontiff, "Does Academic Research Destroy Stock Return Predictability?", Journal of Finance (2016)
• Barber and Odean, "Trading Is Hazardous to Your Wealth", Journal of Finance (April 2000)
• Chague, De-Losso and Giovannetti, "Day Trading for a Living?", SSRN working paper (revised June 2020)
• IRS, Topic no. 409, Capital gains and losses (accessed November 10, 2026)

This article is for educational purposes only and is not investment advice. Past performance does not guarantee future results.