4 days ago
24
One of Europe’s largest banks just handed a chunk of its coding workload to AI, and the results are turning heads on the trading floor. ING Groep has been running what it calls “vibe coding” experiments since February, using artificial intelligence to generate functional code from plain-language requests rather than traditional manual programming.
The target: the bank’s electronic trading infrastructure for foreign exchange and credit markets. Tasks that previously consumed days or even weeks of coding and testing now wrap up in hours, according to Simon Bevan, ING’s global head of e-trading.
What vibe coding actually means
Instead of a developer writing thousands of lines of code by hand, they describe what they want in natural language, and an AI model generates the code for them. The concept gained traction following the advocacy of AI researcher Andrej Karpathy, particularly in early 2025.
ING’s quantitative analysis team has adopted the technique as part of its daily workflow. The primary applications fall into two buckets: prototyping and testing for trading systems, and in-house market data visualization. Both are areas where speed matters enormously and where traditional development cycles have long been a bottleneck.
The bank is running its vibe coding through an in-house Anthropic AI model. By keeping the AI deployment internal, ING maintains data privacy, a non-trivial concern when you’re dealing with proprietary trading strategies and market data. The deployments operate under strict supervision, meaning there’s still a human checking the AI’s homework before anything touches a live trading system.
The business case: speed, savings, and strategy
ING’s vibe coding initiative frees its senior developers to focus on more complex, revenue-generating projects. The cost savings are described as substantial, though ING hasn’t put a specific dollar figure on them publicly.
Bevan also highlighted the ability to create tailored trading instruments. Rather than relying on off-the-shelf solutions, ING can now build bespoke tools fitted to its specific trading strategies and market views.
Why third-party tech vendors should be nervous
Vibe coding threatens to undermine the traditional vendor equation. If banks can generate custom tools in hours rather than months, the value proposition of expensive third-party solutions starts to erode.
Bevan anticipates widespread adoption of vibe coding across the banking sector within a year. His warning was blunt: institutions that don’t adapt risk falling behind those that do.
What investors should watch
Bevan’s emphasis on regulatory adaptation deserves attention. He flagged the pressing need for regulators to keep pace with technological advancements. AI-generated code in trading systems raises legitimate questions about accountability, auditability, and systemic risk.
Bevan also stressed the importance of reliability in trading systems to mitigate risks. Trading infrastructure demands near-perfect uptime and predictable behavior. AI-generated code, even under human supervision, introduces a new category of operational risk that banks and regulators are still learning to manage.
ING’s vibe coding initiative is entirely focused on conventional electronic trading, specifically FX and credit markets. None of the reporting indicated any association with cryptocurrency initiatives, underlining that this venture specifically focuses on enhancing conventional electronic trading tools.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
English (US) ·