Is forex trading still profitable in 2026? The honest math after spread compression and prop-firm collapse

The short answer

Retail forex trading is technically profitable for a small minority of operators in 2026 — the same way poker is profitable for the top 5% of players, with most of the rest funding their win rate. The category isn’t dead; the way it was sold is. The “trade two hours in the morning, scale to $10K/mo” version of retail forex never worked at scale, and the 2022-2024 wave of spread compression, prop firm collapses, and AI-driven institutional execution narrowed the margin further.

If you came here looking for “is forex passive income”: no, it’s not, never was, and the operators who actually make money treat it as a job with capital risk attached — not a stream that runs in the background. If you came here willing to update toward what shape of forex (if any) still works for a retail operator: this is the honest map.

The retail-forex baseline

Every EU-regulated broker is required to publish the percentage of retail accounts that lose money. The disclosed numbers cluster between 70% and 82% depending on the broker and instrument mix. eToro’s own disclosed CFD losses sit around 76%. Plus500 publishes similar. IG Group’s UK-disclosed number for CFD retail clients has hovered around 75% for years. These numbers measure the percentage of accounts that lose money over the disclosed period — typically a quarter or a year. The picture for “the percentage of retail accounts that are still profitable after three years” is substantially worse; most accounts that survive a year are still net negative by month 36.

The math is harsher when you account for survivorship. The 18-25% that aren’t reported as “losing” in a given period include a mix of: short-term winners who haven’t blown up yet, traders who deposited and stopped trading, and accounts that broke even with a small skim. The genuinely net-profitable retail traders — defined as “withdrew more than they deposited across all positions over a multi-year window” — are in the low single digits. The published 25-30% “not losing” number is the wrong benchmark; the realistic profitable-retail-trader figure is closer to 3-5%.

This was true in 2018. It was true in 2022. It’s true in 2026. The category’s win-rate floor hasn’t materially improved with platform improvements, with the introduction of copy-trading, with social-trading networks, or with the AI tooling that arrived in 2023-2024. If anything it tightened.

What changed 2022-2024

Spread compression on majors. EUR/USD spreads at retail brokers compressed from 1.0-1.5 pips to 0.4-0.8 pips over the past three years. Sounds good — until you realize spread compression means the broker’s own margin compressed too, which forced the rise of “raw spread + commission” pricing models that look cheaper headline-wise but produce similar all-in costs once you account for the commission per round-trip. The structural cost of retail forex didn’t drop; it migrated into commission lines that are harder to evaluate in marketing copy.

Institutional execution went fully AI-driven. The other side of every retail forex trade is now executed by algorithmic systems that have access to order book data the retail trader doesn’t. The information asymmetry that always existed in forex widened materially between 2022 and 2024 as institutional desks deployed reinforcement-learning systems for liquidity provision and order matching. The retail trader’s edge — already narrow against floor traders in the 2010s — is now competing with systems that update their pricing decisions on a millisecond cadence using order flow signals the retail trader cannot see.

The prop firm wave collapsed. From 2021 to 2023, hundreds of “funded trader” prop firms appeared offering to fund traders who passed a paid evaluation. The economics were structurally elegant for the firms (most traders failed evaluation; firms kept the fee; the small number who passed traded simulated accounts that were sometimes never actually funded). Through 2023-2024, MetaQuotes restricted MT4/MT5 licenses for prop firms, regulators in the US and EU started classifying some prop firm operations as unregistered securities offerings, and a wave of high-profile collapses (My Forex Funds, Skilled Funded Traders, dozens of smaller firms) hit the category. The survivors — FTMO, FundedNext, The5%ers — operate more conservatively than the 2022 wave, with smaller payout sizes and stricter evaluation rules. For a retail trader, the realistic outcome at a prop firm in 2026 is “you might pass evaluation, you might earn $1-5K in a profitable month, and you’ll definitely give back 40-60% of profit to the firm.” It’s a job structure, not a passive income structure.

Copy-trading saturated and de-prioritized. eToro’s social-trading layer worked in 2017-2020 as a “find a profitable trader, copy them” mechanism. By 2026 the platform’s leading copy-traders carry track records that are heavily filtered through survivorship — most operators with long-track records who actually compounded retired or stopped publishing; the highly-followed accounts visible today are either short-track-record momentum traders or marketing accounts subsidized by the platform. The category’s promise — “you don’t need to learn trading, just copy the winners” — quietly stopped reproducing.

Signal services compressed in value. Telegram, Discord, and (since 2023) AI-generated signal services produce more “trade signals” than ever, of lower aggregate quality than ever. The structural issue is the same as content marketing: production is free, distribution is the moat. The good signal providers (rare) charge $200-500/mo and produce ~55-58% win rates with positive expected value over a long sample. The mass-market $50-150/mo services produce noise.

What actually still works (for the small minority)

Boring discretionary trading in a specific subniche. A retail trader who picks two or three specific instruments (e.g. GBP/JPY during London session and EUR/USD during NY open), develops a narrow setup over 6-18 months of paper and small-size live trading, and disciplines themselves to <10 trades per week can clear $500-3K/mo on $20-50K capital. The annualized returns are 15-30% — better than dividend ETFs, vastly worse than the marketing materials promised. The job is largely “wait, don’t trade, occasionally trade, log everything, refine.”

Prop firm grinding (treated as a job). If you can pass the evaluation phase reliably (top-decile skill, several attempts, $200-1,000 in evaluation fees), the funded phase can produce $1-5K/mo on $50K-200K simulated capital. You give back 40-60% to the firm. The total income matches a mid-level remote engineering job, with vastly more day-to-day stress and worse healthcare downstream. People do it because they like the work, not because the hourly return beats alternatives.

Algorithmic execution on top of a manual edge. A small subset of operators code their own execution layer around a manually-developed edge — basically automating their setup-recognition and order placement to remove emotional drift. This is closer to “small fund operator” than “retail trader.” The capital scales (you can run $100K-1M without changing strategy), but the development time is real (300-600 hours of programming and backtesting before live), and you’re now competing with bigger players on infrastructure.

Trading-adjacent income. Selling indicators on TradingView (covered in our TradingView indicator subscription idea), running a paid signals service if you have a verified edge, writing newsletter content for traders, or building tooling for the small subset of profitable traders. This isn’t trading — it’s serving traders, which has very different unit economics and a vastly larger viable population.

What does NOT work in 2026

• High-leverage scalping with no edge. The most popular retail playbook (50-200x leverage, 30-100 trades per day, “scale your account fast”). Mathematically, traders running this without an edge have a >95% probability of blowout within 6 months. The leverage compresses the time-to-blowout; it doesn’t change the expected value.
• Copying YouTube traders. The accounts most aggressively marketing their results are almost universally either trading with the broker that pays them affiliate commissions on referred new accounts (so they make money whether or not their copy-traders win) or running courses (where the course is the real income, the trading P&L is the marketing material).
• Signal services bought from Telegram or Discord groups. The base rates are bad. Most retail signal services produce <50% win rates over long samples. The 50%+ ones charge premium prices because their edge is real and limited.
• Trading bots sold as “fully automated $200/day.” This is the same category fraud as 2008-2015 — “trade alone in a black box, make money sleeping.” The bots that work are either coded by the trader to their own edge or institutional systems on order book infrastructure retail can’t access. Anything sold for $97 or $497 with claims of “fully automated” is not the math you think it is.
• Demo-account skill misleading you. Two months of demo-account profitability does not generalize. The execution costs, slippage, emotional drift, and psychological pressure of live trading produce typical retail-trader account profitability that’s 30-50% worse than the same operator on demo. Most operators graduate from demo profitable, blow up live, and either quit or repeat the cycle.

Forex is not passive income (even when it works)

This is the conceptual error that mis-shelved retail forex on most passive-income maps. The marketing language — “trade from your laptop, work two hours, make money sleeping” — does not describe the actual operator experience even at the top decile. Profitable retail traders:

• Spend 10-25 hours/week on chart analysis, news, journaling, and trade execution.
• Sit through high-stress drawdown periods that materially affect sleep and household relationships.
• Reinvest a large fraction of profits in larger position sizes rather than withdrawing — because compounding capital is the only way the absolute dollar return scales.
• Have a 12-36 month “investment in becoming profitable” period where the activity is net negative on a P&L basis but positive on a skill-acquisition basis.

That shape is “small trading business operator,” not “passive-income recipient.” It does not belong in the same conceptual bucket as a dividend ETF or a P2P lending portfolio. If you map your time honestly, the realistic hourly return of profitable retail forex (at the top 5% level) clusters at $30-100/hour for someone clearing $3-10K/mo on $50-150K capital. That’s an OK return, but it’s not “passive” by any honest measurement.

Three rules if you’re still drawn to it

1. Do the math on whether trading is your shortest path to any income outcome. If your goal is “$3K/mo passive within 24 months,” the realistic forex math is: 12-24 month investment in skill, $20-50K capital risk, top-decile execution required, ongoing 10-25 hours/week. For most people, the same goal is reachable faster and more reliably via niche affiliate sites, buying a small SaaS, dividend stock compounding, or a small POD shop with $5-15K capital. Forex sits below the alternatives on expected-return-per-hour for most operators.

2. If you commit, treat it as a job and use an EU-regulated broker. Pick eToro, IG, Plus500, or a similar regulated broker. Stay away from offshore brokers, “international” derivatives platforms, and anyone offering 500x leverage. The regulated brokers’ published statistics are the floor you’re operating against; offshore brokers’ marketing-driven statistics are not real numbers.

3. Use position sizing that survives 10 consecutive losers. Most blowouts come from position sizing that “looked aggressive but reasonable” until the run of losses arrived. Sizing each trade at 0.5-1.5% of account equity (not the leveraged exposure — the actual risk per trade) is the operational floor that keeps you alive long enough to discover whether you have an edge. Traders who size at 5-10% per trade compound out of business inside 12-18 months even with positive expected value.

The bottom line

Forex in 2026 is profitable for the same small minority of retail operators it was profitable for in 2018 — the 3-5% with discipline, narrow setups, regulated brokers, real position sizing, and the patience to spend 12-36 months breaking even before clearing real income. It is not, and never was, passive income. Most other categories on this site offer better expected-return-per-hour for the median operator with $5-50K of capital and 10-15 hours/week.

If you read this article and felt the urge to start trading anyway, that’s useful information about your decision-making — most successful retail traders started despite reading articles like this one, not because of them. If you read it and felt the urge to look at our trading-adjacent ideas instead, that’s also useful information. Pick the bottleneck, pick the category, and stop chasing the headline yield.