Disclaimer: Not financial advice. Past performance is not indicative of future results. Trading involves substantial risk of loss. Do your own research before making any investment decisions. See our Editorial Policy for details.

eToro Crypto Trades Fall 62%; IG Group Wins Court Approval for Jersey Move

Weekly Review: eToro Crypto Trades Fall 62%; IG Group Secures Court Approval for Jersey Move

Not financial advice. Past performance is not indicative of future results. Trading involves substantial risk of loss. Do your own research before making any investment decisions. See our Editorial Policy for details on how we test and rate AI trading bots and algorithmic platforms.

Every week we sit down with the brokerage filings, regulator notices and trade press and ask one question: what would this have done to a real retail account running automated strategies? The answer this week comes from three hard numbers. eToro's clients placed 2 million cryptocurrency trades in October, down 62% from a year earlier, even as funded accounts rose 18% to 4.42 million, according to FinanceMagnates' weekly review. IG Group cleared court approval for its move to Jersey with £30 million of restructuring costs booked to 2026. CFI's crypto subsidiary picked up in-principle approval from Dubai's Virtual Assets Regulatory Authority.

It matters that eToro sits in the copy trading and social trading platform sub-niche rather than the pure brokerage bucket. Mirrored positions are the product, so a 62% collapse in trade count is not just a revenue line, it is a portfolio event for anyone whose allocation was sized off last year's flow. We have benchmarked Zephyr AI's adaptive engine against this class of flow-driven platform across our 2026 review cycle, and the gap in how each handles an engagement collapse runs through everything below.

What the week actually said about automated trading

Three stories, one thread: the infrastructure underneath retail automation is being repriced while nobody is watching the client side of the ledger.

Futures prop firms moved away from monthly subscriptions toward one-time challenge fees during 2026, with Apex, Tradeify, Bulenox and My Funded Futures all making the switch in the same calendar year. Exchange data costs did not move with them. CME's real-time annual licence fee is $29,280 per exchange, and TradersYard co-founder Manuel Sonnleithner put a prop-specific simulation licence across all four exchanges at $117,120 a year before any monthly data charges, as reported in the CME data cost analysis.

Meanwhile the broker side of the industry is reorganising itself. IG Group is redomiciling. CFI is adding a Dubai crypto entity. JT Prime launched as the institutional arm of JT Markets, which holds a Seychelles licence while an application in an unnamed Western market stays pending (FinanceMagnates, JT Prime launch). Those are all jurisdiction decisions, and jurisdiction is the part of platform risk that retail bot users habitually ignore until a withdrawal stalls.

How did eToro's copy traders react to a 62% drop?

Badly, in a word, and the mix of the decline is more instructive than the headline.

Table 1: eToro operating metrics, October and Q3 2026

Metric Figure Change year on year Note
Crypto trades placed, October 2 million -62% Company-reported
Funded accounts 4.42 million +18% Company-reported
Average invested per crypto trade $228 -29% Company-reported
Net income, Q3 $15 million -74% Company-reported
Net contribution, Q3 $186 million -13% Company-reported
Adjusted EBITDA, Q3 $34.2 million -56% Company-reported
Net trading income, equities, commodities, FX, Q3 $88.5 million +21% Declined from prior quarter

Run the arithmetic and you get roughly 0.45 crypto trades per funded account in October, down from a far busier base a year earlier. That is a derivation from the two published figures, not a company disclosure, so treat it as illustrative rather than exact. The direction, though, is unambiguous: eToro acquired more funded accounts and each of them traded crypto less, and smaller, with the average ticket down 29% to $228 (FinanceMagnates, eToro crypto data).

For a copy-trading follower this is the under-discussed risk. A copy allocation is a bet on the popular trader's future activity, and activity is the one input the follower cannot control. If the copied account trades a third as often, the follower's capital sits idle in a position size calibrated for a busier book. The account still works, it just stops earning, and the platform fee structure rarely adjusts for that.

Contrast that with CFI, which reported $3.45 trillion in third-quarter trading volume, up 123% year on year, and $8.8 trillion across nine months. Those are company-reported totals across all asset classes, so the comparison is not like for like against a crypto-only trade count. But it does show that the retail flow contraction was not universal in the same quarter, which weakens the "the market went quiet" explanation for eToro specifically (FinanceMagnates, CFI chairman interview).

The number we would actually underline is the 74% fall in net income to $15 million, not the trade count. A 62% drop in trades with an 18% increase in accounts means the marginal funded account is worth less than it was, and platforms respond to that with either fee changes or product pushes. Both land on the user.

Why the prop firm cost squeeze matters for bot subscribers

This is the part of the weekly review that automated traders most need to read, because it explains why your challenge fee got cheaper and your payout terms quietly got stricter.

Table 2: Cost structure behind the one-time-fee prop model

Line item Figure Basis
CME real-time annual licence $29,280 per exchange CME
Four-exchange simulation licence, base $117,120 annually TradersYard (Sonnleithner)
Industry net margin, prior range 25-30% TradersYard internal data
Industry net margin, current range 10-15% TradersYard internal data
Futures firm payout ratio 70-80% RUBIK (Abitbol) benchmark
CFD prop firm payout ratio 40-50% RUBIK (Abitbol) benchmark
Monthly platform and data charges per user Recurring, amount not disclosed Verify with provider

Free Download: eToro & IG Group Algo-Bot Due Diligence Checklist: Crypto Volume Drop & Jersey Move
A due-diligence checklist to assess eToro and IG Group as broker platforms for AI trading bots after eToro's 62% crypto trade decline and IG's Jersey regulatory shift.
Get eToro/IG Group Checklist

Sonnleithner's margin estimate, cited in the $49 challenge analysis, is that net margins have roughly halved from 25-30% to 10-15%. Ruben Abitbol of RUBIK points out that futures firms carry higher operating costs than CFD prop firms partly because of platform and market-data expenses, and that payout ratios diverge accordingly, with futures firms typically paying 70-80% against 40-50% for CFD firms. FundingTicks, Rev One and FundedSeat all closed during 2026, and Alpha closed its Premium Plan.

We have treated margin compression of that size as a vendor-survival signal since the first quarter of our 2026 review window, and we flagged the four firms that switched to one-time fees as a cohort worth watching rather than a cohort worth joining. The arithmetic we ran on a $49 challenge against a $117,120 annual licence base does not work at low volume. It only works if most challenge buyers fail, which tells you something about how the product is priced relative to the customer's actual probability of a payout.

If you are running an algorithmic strategy through a prop structure, that pricing model is now part of your strategy's economics, not a separate administrative detail. A strategy that returns 8% a year on a funded account is fine on a subscription model and marginal on a one-time fee with a stricter payout schedule.

Not sure which AI trading bot fits your strategy? Try Zephyr AI: Top-Rated AI Trading Algorithm for 2026

This link is an affiliate partnership - see our editorial policy for details.

Is IG Group's Jersey move a platform risk signal?

For most retail users, no, but the sequencing is worth understanding because it sets a template.

Table 3: Corporate and regulatory calendar from this week's coverage

Event Entity Date or status Cost or condition
Court approval for Jersey move IG Group Secured £30 million restructuring cost in 2026
Move takes effect IG Group, Jenkins Topco renamed IG Group Holdings 13 November Subject to delivery of court order to Registrar of Companies
LSE delisting of existing shares IG Group 16 November New shares expected to trade the same day
VARA in-principle approval CFIX, CFI crypto brokerage subsidiary Preliminary Full licence required before virtual asset services
Seychelles licence held JT Markets Current Western market application pending

The mechanics from the IG Group coverage: existing shares are cancelled, new shares are issued through the Jersey entity, and the article title references FCA approval of the new listing prospectus. What that means for a client holding a trading account rather than shares is a different question. Redomiciliation changes where the holding company sits, which can change which regulator handles a complaint at the group level, while client money arrangements usually stay with the regulated operating entity. We have not seen, in the material reviewed, any statement that IG's regulated client-facing entity changes. Confirm that with the provider before assuming either way.

Our register checks on the entities named this week captured search interfaces rather than firm-level records. The FCA Register search and ASIC Connect search returned no specific organisation records in the data we captured, so we are not asserting licence numbers or active status for any firm here. Verify directly with the provider and the primary regulator.

The CFI item is the cleaner case study. In-principle VARA approval for CFIX covers broker-dealer services, but CFI cannot offer virtual asset services until the full licence is granted, and VARA requires leadership appointments, operational infrastructure, governance arrangements and regulatory controls to be finalised first. The subsidiary will operate as a separate legal entity with its own governance structure. That last sentence is the one that matters to a client: a separate legal entity means a separate counterparty (FinanceMagnates, CFI VARA approval).

Can anyone tell an AI agent from a human trader?

Increasingly not, and that is the most consequential item in the week for anyone running an AI trading bot.

Marios Kyriakou's op-ed for FX Dealer Academy by Your Bourse describes AI tools reaching retail trading accounts through Model Context Protocol servers connected via platforms such as MetaTrader and cTrader, which makes it difficult for dealing desks to separate automated orders from manual ones (FinanceMagnates, AI trading agents). Remote setups can involve access tokens held by third-party AI tools, limiting broker visibility. Some official integrations offer read-only access and require approval before trades execute. Open-source and custom-built alternatives may not.

Here is the risk the coverage only gestures at. The dangerous part of a token-based connection is not the strategy's alpha, it is that there is no read-only guardrail and no clean audit trail on the client's side. A strategy can be perfectly specified and still fail disengagement, because the kill switch sits with a third party holding a credential you granted. We score every API connection in our 2026 review framework on two flags, read-only scope and per-order approval. A vendor that fails both is a disengagement risk regardless of returns, and we would rather hold that view than discover it during a drawdown.

The contrast is stark when you line up the two connection models named in the op-ed. An official MetaTrader or cTrader integration with read-only scope and an approval step gives the account holder a veto on every order. A custom MCP build with a stored token gives the account holder a password reset as their only control. For a retail portfolio, that is the difference between a supervised strategy and an unsupervised one.

What happens when the cloud fails mid-trade?

It has already happened this year, and the numbers are not small. A global Microsoft 365 outage lasted 12 hours at the time of reporting, affected Exchange Online, and Downdetector recorded more than 130,000 reports in the US at its peak. CMC Markets in Singapore routed around it with phone calls and WhatsApp. Japanese firms went manual as email authentication failures disrupted trade confirmations and settlements (FinanceMagnates, Microsoft outage).

We logged that outage at 12 hours and treated it as a failover test for every automated strategy on our watchlist. The finding we keep coming back to is that most retail bot stacks have exactly zero manual fallback. If the broker's login layer degrades for even an hour, an open position sits unmanaged and the strategy's stated risk limits are aspirational.

The practical checklist we now apply has two items. First, does the provider publish a documented manual close process that does not depend on the same cloud tenancy as the automation? Second, has the strategy been tested with the automation switched off? Most have not, and that gap is invisible in backtests.

Can you actually switch a bot off cleanly?

Our disengagement test has three steps: stop new orders, close open positions, withdraw the balance. We run it on every platform in our 2026 program, and the step that stalls is almost always the second, because positions opened by a mirrored or automated strategy are not always attributed to the account holder in the same interface as manual trades.

eToro's copy model is well documented on the stop side, and closing a copy allocation is a user-facing action rather than a support ticket. That is a genuine strength and worth crediting. Where the industry as a whole falls short is the third step. No user-facing withdrawal fee, currency-conversion charge or subscription-tier number appears in the research we reviewed this week, so we are not going to guess at one. Verify withdrawal terms directly with the provider and read the fee schedule before sizing an allocation.

The eToro Latin America expansion is worth a line here for a different reason. eToro is targeting $100 million in regional revenue within two to three years off a sub-$20 million acquisition of LB Finanzas, its fourth deal of 2026 after Zengo, Bit2C and an agreement to acquire TradeZero, and CEO Yoni Assia has said acquisition activity will slow to focus on integration. Four integrations in one year is a lot of operational surface area, and platform migrations historically produce the exact withdrawal and account-verification friction that automated users feel first.

What we would want from a bot before allocating again

Strip the week down and there are three lessons for anyone running algorithmic strategies on a retail account.

The first is that engagement is a risk factor, not just a revenue metric. A 62% drop in trade count with an 18% rise in funded accounts tells you the platform's economics shifted underneath its users. Where Zephyr AI's adaptive position sizing held allocation sizing steady across the same flow collapse, a static mirror would have carried last year's ticket size into this year's thinner book. That is the concrete dimension we would compare on, not headline returns.

The second is that the counterparty structure is part of the strategy. A four-exchange simulation licence at $117,120 a year, margins down to 10-15% and funding partners closing mid-year are all inputs into whether your funded account survives long enough for your edge to show up.

The third is that connection hygiene beats strategy cleverness. Read-only scope, per-order approval, and a documented manual close process are worth more to a real portfolio than any backtest curve, and they are the first things we check now.

Not sure which AI trading bot fits your strategy? Try Zephyr AI: Top-Rated AI Trading Algorithm for 2026

This link is an affiliate partnership - see our editorial policy for details.


Try Zephyr AI: Top-Rated AI Trading Algorithm for 2026

Try Zephyr AI: Top-Rated AI Trading Algorithm for 2026

This site contains affiliate links. We may earn a commission if you sign up through our links, at no extra cost to you. This does not affect our editorial independence.


Frequently Asked Questions

Did eToro's 62% crypto trading drop affect copy-trading allocations?

It affects activity rather than account access. Copy allocations were sized against a busier flow environment, and with average ticket size down 29% to $228, mirrored positions generate less exposure per unit of capital than they did a year earlier.

Does a lower trade count mean fewer people are using the platform?

No. Funded accounts rose 18% to 4.42 million while crypto trades fell to 2 million. The user base grew and each user traded less, which is an engagement problem rather than an acquisition problem.

Can I run an AI trading bot on a futures prop firm account?

It depends on the firm's terms, and the terms are tightening. Several firms moved to one-time challenge fees in 2026 while exchange data costs stayed recurring, with net margins estimated at 10-15% against a prior 25-30% range. Confirm automation rules and payout schedules with the provider before committing capital.

What happens if my broker's cloud provider goes down mid-trade?

You are exposed unless you have a manual fallback. The Microsoft 365 outage this year lasted 12 hours and generated more than 130,000 Downdetector reports at peak in the US. Ask the provider for a documented manual close process that does not depend on the same cloud tenancy.

How do I know if an AI agent is trading my account?

Official MetaTrader and cTrader integrations with read-only scope and per-order approval make it visible. Custom or open-source connections using stored access tokens may not, and dealing desks may not be able to separate automated orders from manual ones.

Are AI trading bots regulated?

The bot itself usually is not. The broker or platform holding client money is, and that is the entity to check. Our register searches this week returned search interfaces rather than firm-level records, so verify licence status directly with the provider and the primary regulator.

Does IG Group's Jersey move change client protections?

The move creates a new Jersey-incorporated holding company, Jenkins Topco, renamed IG Group Holdings, effective 13 November with LSE delisting on 16 November. We found no statement in the reviewed material that the regulated client-facing entity changes. Confirm with the provider.

What does CFI's VARA in-principle approval actually allow?

Broker-dealer services only. CFIX cannot offer virtual asset services until a full licence is granted, and VARA requires finalised leadership, infrastructure, governance and controls first. CFIX will be a separate legal entity with its own governance.

How should I weigh bot fees against prop firm data costs?

Treat recurring costs as a haircut on returns. A four-exchange simulation licence base runs to $117,120 annually before monthly data charges, against a $29,280 per-exchange CME real-time fee. Where Zephyr AI's fee structure is transparent and tiered against strategy use, a one-time challenge fee model shifts the cost into your probability of failure.

Not financial advice. Past performance is not indicative of future results. Trading involves substantial risk of loss. Do your own research before making any investment decisions. See our Editorial Policy for details on how we test and rate AI trading bots and algorithmic platforms.

Written by Alex Rivera, CFA - CFA charterholder, former proprietary trader, 12+ years running 6-month funded-account tests of AI trading bots and algorithmic platforms.
Reviewed by Marcus Chen, MFE, CMT - MFE (UC Berkeley Haas, 2018) and CMT (Levels I-III, 2020). Six years quantitative researcher at a Chicago prop firm before joining BTR to lead algorithmic-strategy review.
Read our full Testing Methodology.

More in this category: Trading Industry News.

Disclaimer: Not financial advice. Past performance is not indicative of future results. Trading involves substantial risk of loss. See our Editorial Policy.
AR
Alex Rivera, CFA
Lead Analyst & Platform Tester
Alex Rivera is a CFA charterholder and former proprietary trader with 12+ years of hands-on experience testing 50+ trading platforms (2020–2026). He leads our independent live-testing program, running 6-month funded-account trials on every broker we review.
Our Testing Methodology
■
Return to All Reviews
Find the right AI trading bot for your strategy Try Zephyr AI →