This article is authored by John Stead, Director of Sales Specialists & Marketing at smartTrade
FX Markets Europe 2025 concluded last week in London. While the panels covered everything from macroeconomics to regulation, the real story on the floor was about infrastructure. The message is clear: the plumbing of the FX market is under immense strain, and the gap between legacy technology and future requirements is widening.
Having spoken on the New Rules, New Rails panel and listened to the deep-dives throughout the day, here is my view on the architectural challenges facing the market—and how we solve them.
1. The Capacity Challenge: “The Brakes Are Off”
If you believe your current market data infrastructure is adequate, you need to look at the numbers shared by Marc Hinken (NatWest). He noted that NatWest consumed 2.5 times the amount of market data last year compared to a decade ago. Even more telling from a capacity planning perspective: in April alone (during the “Liberation Day” volatility), they processed 50% of the total market data volume of 2014.
Hinken explained that this isn’t just “more trading”; it is structural. Historically, primary markets acted as a natural brake with 50ms updates and half-pip precision. Today, with millisecond updates and quarter-pip precision, that brake is gone. As Jeremy Smart (XTX) rightly pointed out, much of the banking technology stack is decades old and simply not architected for this continuous firehose of data.
ST View: If your tech stack is built for linear growth, it will fail. You need an architecture designed for exponential throughput to handle these peaks without adding latency to your decision logic.
2. The Liquidity “Mirage” & The Flight to Quality
The explosion of data has created a “liquidity mirage” where visible prices on public order books are a poor proxy for actual depth. Ayesa Latif (Citi) revealed a staggering metric: during volatility events, Citi’s internalisation rates can hit 1,000% of visible liquidity.
We also heard from Ralf Donner (Goldman Sachs) about “phantom liquidity” on anonymous ECNs—participants that vanish the moment stress enters the market. Consequently, when volatility spikes, we see a massive flight to quality. Latif noted that rejection rates on aggregators spike during stress, causing clients to flood back to Single Dealer Platforms (SDPs) where they have a direct line to a bank’s internal risk warehouse.
ST View: You can no longer rely on simple aggregation. The “intelligence” of the execution stack matters more than ever. Sell-side firms need sophisticated engines capable of warehousing and internalising flow dynamically. At smartTrade, we see this shift first-hand: clients are demanding engines that manage this logic in real-time, rather than just acting as pipes to external venues.
3. The Interoperability Challenge: Ledger vs. Blockchain
On the stablecoin panel, I was joined by Sabih Behzad (Deutsche Bank) and Emma Lovett (JPMorgan) and moderated by the excellent Pierre Pourquery (Marceau Partners).
The consensus? We are moving from pilot mode to production.
ST View: You cannot simply treat a stablecoin like another fiat currency pair. We are witnessing a fundamental architectural shift from Accounts (database ledgers) to Wallets (blockchain interaction). The winners will not be those running parallel systems; the winners will be those who can integrate digital asset execution and settlement directly alongside Spot and Swaps in a single, unified stack. smartTrade with clients in crypto and fiat currencies for many years is of course the ideal partner to help you navigate this hybrid market!
4. The Digitisation Challenge: Swaps
It was excellent to hear from Simon Wilson-Taylor (FXswapX). He highlighted a massive inefficiency: in a $4 trillion market, 79% of inter-dealer volume is still traded via voice.
Simon made a very compelling case for innovation in the inter-dealer space. Specifically, he spoke about moving away from the model of monetising data, which often discourages banks from sharing their best price. His approach—creating a mid-rate using encrypted data to prevent information leakage—is exactly the sort of privacy-focused innovation the swap market needs.
ST View: We see many clients coming to us for help with modernising their swaps price discovery and distribution infrastructure. Initiatives such as this, which increase digitisation in the inter-dealer market, can only help bring more efficiency to the market—ultimately benefiting the underlying clients on whom we all depend.
5. The Adaptation Challenge: AI in the Stack
Finally, we saw practical implementations of AI from both Quoniam Asset Management and Bank of America. We are moving beyond “AI as a buzzword” to “AI as a component.”
ST View: We are seeing Large Language Models (LLMs) integrated into sentiment analysis workflows , and Deep Reinforcement Learning being deployed to adapt execution algorithms in real-time based on fill probability. This requires a tech stack open enough to ingest these models and robust enough to govern them in production.
Final Thoughts
The common thread across all these talks—from the “black screens” in Swaps mentioned by Hinrich Wilhelm Paul (Commerzbank) to the data surges at NatWest —is that “good enough” legacy technology is becoming a liability. The market is being re-platformed, and the complexity is increasing. If you would like to discuss any of these topics or to hear how smartTrade is helping clients address general infrastructure challenges please do get in touch.