A new report from HC Group and FT Longitude, published in April, found that AI is expected to deliver an average additional 3.34% uplift in trading P&L by next year. For firms operating on thin margins across volatile markets, that is no mean feat.
The same research also found that 98% of commodity trading executives are already seeing measurable value from AI. But when we asked the room at CTW Europe last month, only around 15% of attendees said they felt ahead on AI.
That gap is the most interesting thing happening in commodity trading right now. And it will be at the centre of the conversation at Commodity Trading Week Americas, the event that brings together 650+ senior leaders from across trading, risk, technology, and operations in Stamford, Connecticut next week.
The paradox hiding in the data
The problem is clearly not in getting started with AI. It is the distance between AI delivering value in targeted use cases, and becoming a genuinely integrated capability whose value compounds over time.
When you look at it through that lens, the 15% figure from CTW Europe starts to make more sense. Trading firms are not being slow to adopt AI. Many have run pilots, explored tools, and seen early results. But they feel behind because they are measuring themselves against what AI could eventually do, not against where their competitors actually are today.
Why data keeps coming up
The HC Group report identified data quality, access, and organisation as the single biggest barrier to scaling AI adoption. This finding is consistent with almost every conversation we have had over the past year. The tools exist. The intention is there. What stalls progress is the underlying data environment: fragmented systems, manual inputs, and inconsistent structures.
AI cannot build reliable decision support on unreliable foundations. If the data going in is incomplete or already out of date, the outputs will reflect that, and firms will either distrust the results or, worse, act on them anyway.
This is not a criticism of the firms grappling with their data. Commodity trading data is genuinely complex, and trading companies have been working to solve that complexity for decades. You are dealing with physical and financial positions, bespoke contract terms, multi-leg logistics, and pricing across dozens of variables, often managed across systems that were not designed to talk to each other. Getting that data into a state where AI can work with it is a real undertaking.
But it is also what separates the firms who are seeing that real P&L uplift. So the question worth asking ahead of CTW Americas is: “Is our AI actually working with our data, or around it?”
Bringing traders and experts together
Gen10’s Brett Anderhub will be leading one of the Audience Interactive Discussion tables at CTW Americas on Thursday 18th June. One of the things that makes the session worth attending is the format itself; a roundtable structure that puts the audience in conversation with each other.
The best insights often come from the exchange between someone who has just navigated a data migration or technology implementation and someone who is about to start one. That kind of conversation does not happen easily in a panel format, but the roundtable structure at CTW Americas is designed to facilitate it. Brett will lead two rounds of conversation around AI in commodities, before feeding back your insights to the room.
Where the useful conversations will be
If you are attending CTW Americas and the question on your mind is some version of “we know AI matters, but we cannot seem to get past the data problem”, that is something we can help with. Join Brett at the roundtable on Thursday, or at the Gen10 stand throughout both days.
If you would like to arrange time in advance, whether that is a meeting with Brett, or a live CommOS demo built around your own contracts and commodities, get in touch and we will set something up before the event.
