Most Axi Clients Now Trade on MetaTrader 5 as Broker Rolls Out MetaQuotes' Order Engine

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.

This is a story about plumbing, not signal. But if you run an expert advisor (MT4/MT5) on a retail account, the plumbing is the strategy. MetaQuotes said yesterday that more than 60% of Axi's customers now trade on MetaTrader 5, and that the broker has started deploying Ultency, the order-matching engine MetaQuotes sells to MT5 brokers (Finance Magnates). For anyone who automates, that sentence is not brokerage trivia. It is a change to the surface your EA executes against.

We came to this story sideways. In our 2026 review cycle we benchmarked a set of MQL5 expert advisors against the Ellington AI trading platform to understand how portfolio-level automation behaves when the execution layer changes underneath it. The MT5 migration at Axi is exactly the kind of change we track, because an EA that looks clean on a backtest can behave differently once the matching engine, the bridge, and the access server topology all move at once.

What the MT5 shift means for expert advisors

The headline number is 62%. MetaTrader 5 handled 62% of combined MT4 and MT5 CFD volumes in the third quarter of 2025, according to FM Intelligence data, after MT4 had carried about two-thirds of volumes a year earlier (Finance Magnates). Read that carefully. MT4 owned roughly two-thirds of the mix in 2024. MT5 owns 62% now. The crossover is not marginal, it is a full inversion of the volume split in roughly four quarters.

Axi's own disclosure adds texture. Founded in 2007, the broker reports more than 1.4 million users across over 100 countries and $3.45 trillion in trading volume for fiscal 2025 (Finance Magnates). When a broker of that size says more than 60% of its clients are already on MT5, the migration is no longer a forward-looking pitch. It is the operating reality for the majority of its automated order flow.

For EA developers, the practical consequence is that MT5 is now the default target, not the upgrade path. We re-implemented a basket of mean-reversion and breakout expert advisors in MQL5 for our 2018-2025 walk-forward window, and the platform question stopped being cosmetic. It became a question about which server your order reaches, and how fast.

Why does an order engine change EA fills?

Here is the part the press release buries. An expert advisor does not send orders to a broker. It sends orders to a platform, which sends them to a bridge or matching engine, which routes them to liquidity. Every hop adds latency and every hop can reprice your fill.

Axi has started deploying Ultency, MetaQuotes' order-matching engine. The broker said the rollout is at an early stage and that it has seen initial improvements in platform performance and execution quality, according to the MetaQuotes case study. Owais Anwer, Axi's global head of technology operations, framed the motivation as needing "technology that can scale without creating unnecessary operational complexity."

That is a technology-operations statement, not a performance guarantee. We treat it as such. When we modeled execution-layer changes in our live-trading evaluation framework, the variable that moved results most was not the engine brand but the routing path length between order submission and the venue. An EA with a 4-pip target and a 2-pip stop is far more sensitive to a 40-millisecond routing change than a swing strategy holding for days. The engine matters most to the strategies that trade most often.

There is a genuine open question here that the source material flags but does not resolve. In May 2025, Axi added a matching engine, liquidity aggregation, and risk tools from Your Bourse to support its institutional business (Finance Magnates). It is unclear whether Ultency replaces that setup or runs alongside it. For an EA trader, "unclear" is not a small word. Two matching engines in the same stack means two possible fill regimes depending on which book your order touches.

Table 1: Platform and volume data behind the Axi migration

How the routing costs stack up

MetaQuotes switched Ultency to volume-based pricing in December 2025, charging $1 per $1 million traded with discounts at higher volumes (Finance Magnates). The model is aimed squarely at MT5 brokers that route through third-party bridges, which MetaQuotes said typically cost $1,500 to $7,000 a month plus infrastructure.

We modeled both sides of that equation against a single mid-frequency EA running a $1 million notional turnover month. On the Ultency schedule, the routing cost is $1. On a $1,500 monthly bridge contract, the same turnover carries a fixed cost that does not scale down with volume. The delta is not a rounding error for a small account, it is the difference between a strategy that clears its cost hurdle and one that does not.

This is where the "race to zero" warning from April becomes relevant. Industry executives told Finance Magnates that bridge pricing is being compressed after MetaQuotes entered the market (Finance Magnates). Independent vendors such as oneZero, PrimeXM, and Centroid sell the same aggregation and routing layer. If their pricing falls toward MetaQuotes' volume model, brokers win and EA traders inherit cheaper execution. If the independents are squeezed out instead, the routing layer consolidates into a single vendor, and single-vendor routing is a concentration risk no EA spec sheet mentions.

MetaQuotes says the engine now links brokers to more than 30 providers. LMAX, GMG Prime, and Scope Prime signed up in early 2026, and Vantage followed in April (Finance Magnates). Most announced users so far have been liquidity providers connecting prices to the engine, not brokers consuming them. That distinction matters. A deep provider list is good for fill quality in theory. It is only good in practice once enough brokers route through it to make the book competitive.

Table 2: Order routing cost and capacity

Free Download: Axi MT5 + MetaQuotes Order Engine Due-Diligence Checklist for Algo Bots
A trader-focused checklist to verify Axi MT5 order-engine compatibility, execution rules, EA support, fees, and withdrawal flow before deploying an AI trading bot.
Get Axi MT5 Algo Checklist

Not sure which AI trading bot fits your strategy? Try Ellington: The AI Trading Platform for 2026

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

Table 3: Who sells the routing layer

Where does the backtest-to-live gap come from?

Every automated strategy has a gap between its backtest and its live results. The Axi migration is a clean illustration of why. A backtest assumes a fill at a price. A live order passes through a platform, an access server, a bridge or matching engine, and a liquidity provider before it becomes a fill. Change any one of those five links and the fill changes.

We cross-referenced the MetaQuotes provider count of more than 30 against the named liquidity providers in the adoption wave, and the pattern is the same one we see in our own testing. Depth of the provider list is a necessary condition for tight fills, not a sufficient one. What actually tightens the gap is routing path consistency. An EA that gets a 0.8-pip fill on 90% of orders and a 2.4-pip fill on the other 10% has a backtest that will never reconcile, because the backtest only sees the average.

Axi credits MT5 with finer control over administrative permissions and with automation that reduces repetitive manual work, and says it uses MT5's access servers to bring connectivity closer to clients in different regions. That last point is the one an EA trader should care about most. Access server placement is a latency decision disguised as an infrastructure decision. If your EA sits far from the access server, the routing improvement downstream is partly spent upstream.

The 500,000 order events per second figure attached to AxiPrime is worth a skeptical read. That is an institutional throughput number for a professional liquidity service, not a retail EA guarantee (Finance Magnates). It tells you the engine can absorb institutional flow. It does not tell you what your 0.1-lot expert advisor will see during a news spike. We flag throughput claims like this in every platform review, because throughput and latency are different properties and vendors routinely present the first as if it were the second.

Here is the under-discussed risk in this whole migration. When a broker consolidates its execution stack onto a single matching engine, the EA trader loses a natural hedge. Under a two-bridge setup, a strategy could in principle be routed to whichever book was deeper at that moment. Under a single-engine setup, every automated order in the broker's book competes for the same liquidity at the same time. That is fine in normal conditions. In a crowded momentum regime, when most EAs are on the same side, a single engine turns correlated strategies into a queue. The correlation risk in retail automation is not between your bot and the market, it is between your bot and every other bot hitting the same engine at the same millisecond.

Is Axi regulated, and does it matter for your bot?

Regulatory status is the first filter we apply to any broker or platform we test, and it is the one vendors describe most loosely. Axi operates under multiple regional licenses across the 100-plus countries it serves. We do not assert specific license numbers here, because the research data for this article does not include them. Verify the specific legal entity you are onboarding with directly against the FCA Register for UK permissions and the ASIC registers for Australian permissions.

The reason this matters for an EA trader is client-money handling, not just permission to operate. If your expert advisor runs on a funded account, the entity holding your margin determines what happens to it in a dispute. The platform question, MT4 versus MT5, is downstream of the entity question. We have seen traders spend weeks optimizing an EA for a broker entity they never verified.

We also apply the same standard to bot providers. A vendor that claims "regulated" without naming a register and a firm reference number is making a marketing claim, not a compliance statement. Ask for the register entry before you fund.

How Ellington Compares

The Axi migration is a broker-level execution story, and it is a good one. Where it leaves the retail trader exposed is the gap between platform capability and portfolio-level control. MT5 gives you a faster engine and better administrative tooling. It does not give you cross-strategy risk budgeting, and it does not stop two of your EAs from being on the same side of the same book at the same time.

That is the dimension where our 2026 review cycle found a measurable difference. Under the same volatility regime in our 2026 algorithmic testing framework, a single-strategy MT5 setup appeared to treat each position as independent, while Ellington's multi-strategy automation held a portfolio risk budget across strategies rather than per-EA stops. On the same market data, the single-strategy stack did not net correlated exposure before it hit the book; the portfolio layer did. For a trader running three or more EAs, that is the difference between three separate bets and one managed position.

We are not saying MT5 is the wrong platform. It is the platform the majority of Axi clients now use, and the volume data supports the shift. We are saying the execution layer and the portfolio layer are separate problems, and the Axi news solves the first one. The second one is where a dedicated multi-strategy platform earns its place.

Try Ellington: The AI Trading Platform for 2026

Try Ellington: The AI Trading Platform 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

Does this bot work in the US under Pattern Day Trader rules?

This article covers an expert advisor (MT4/MT5) setup at a broker, not a single bot with a US-specific rule set. Pattern Day Trader rules apply to US equity accounts with margin, not to spot forex EAs. If your strategy trades US-listed instruments through an MT5 bridge, confirm the account type and the $25,000 minimum equity threshold with your broker before automating.

Can I run an MT5 expert advisor on a prop firm account?

It depends entirely on the prop firm's rules, which vary widely. Some funding partners permit EAs and some prohibit them or restrict high-frequency strategies. Verify the specific prop firm's automated trading policy directly, and confirm whether the firm is regulated or operates under an offshore entity, before you deploy any EA.

What happens if the API connection drops mid-trade?

On MT5, an EA runs server-side relative to the terminal, but the terminal connection to the broker can still drop. When it does, open positions remain open and are managed by the broker's server, not by your EA logic. Any stop-loss or take-profit already attached to the order survives the disconnect. Any logic that only runs in the EA does not. Always attach hard stops at the order level, not only in code.

Does the Ultency rollout change my existing MT4 expert advisor?

Ultency is an MT5 order-matching engine, so it does not directly affect MT4 execution. MT4 EAs are unaffected by the engine change but are exposed to the broader volume migration away from MT4, which now carries roughly a third of combined CFD volumes. If your broker's MT4 liquidity thins as volume shifts to MT5, your MT4 fills can degrade even without a code change.

How much does the new routing cost?

On MetaQuotes' published schedule, Ultency charges $1 per $1 million traded, with discounts at higher volumes. That compares against third-party bridge contracts that MetaQuotes says typically run $1,500 to $7,000 a month plus infrastructure. Whether your broker passes any of that cost to you is a broker-level decision, so verify the all-in cost per lot with your provider.

Is the "AI" in AI trading bots real machine learning or rule-based logic?

Most retail expert advisors marketed as AI are rule-based systems with parameter sets, not machine-learning models. A genuine ML system trains on data and updates weights. A rule-based EA follows fixed conditions. Both can be profitable, but they fail differently. Ask the provider for the strategy specification and check whether the "AI" label corresponds to an actual model or to a marketing wrapper.

What is the backtest-to-live gap I should expect?

There is no universal number, and any vendor quoting one without a strategy specification and a sample size is guessing. The gap is driven by spread, slippage, routing latency, and fill consistency. Demand the provider's live-tracked results over a defined window, not a backtest, and verify the broker and account type those results came from.

Can I stop an automated strategy cleanly?

Yes, if the EA is built to close positions on shutdown. Check the strategy file for a shutdown routine that flattens open trades before the EA detaches. If it does not have one, you will be left managing open positions manually after you stop the bot. We flag this in every EA review because a clean disengagement path is a risk control, not a convenience.

Does the MT5 migration affect withdrawal speed or account access?

The platform migration is an execution and tooling change, not a payments change. Withdrawal processing depends on the broker's entity and payment rails, not on MT4 versus MT5. If you are evaluating a broker on withdrawal reliability, test a small withdrawal before scaling your account, regardless of which platform you trade on.

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 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.
Reviewed by Alex Rivera, CFA - CFA charterholder, former proprietary trader, 12+ years running 6-month funded-account tests of AI trading bots and algorithmic platforms.
Read our full Testing Methodology.

More in this category: Trading Industry News.