Reference Data Intelligence

Data that actually makes sense.

MetalsData.AI is built on a governed, versioned Knowledge Engine that shows how physical metals exposure, hedging instruments, benchmarks and time relate to each other — so your data isn't just numbers, it's understood.

Governed · Authoritative · Complete

What those three words actually mean

Not slogans, but design principles. Here's what each one commits us to.

1

Governed

Our Knowledge Engine isn't hand-maintained. When a source's reference data changes, the Engine is designed to rebuild from it — with every change logged and traceable, not a black box you have to trust blindly.

2

Authoritative

Modelled directly against how Exchanges and Price Reporting Agencies actually define the market — not a second-hand simplification that flattens everything into separate siloed datasets.

3

Complete

One structure that connects physical exposure, hedging instruments, benchmarks and time — so you can see how a data point relates to the rest of the market, not just where it sits in isolation. When we say a dataset is complete, we mean 100% complete — nothing left out.

Why this is different

Most providers just sell you prices. We ship the structure with the prices, so you know what they actually mean.

A Knowledge Engine, not a spreadsheet

Every entity — Futures contract, benchmark, warehouse location — is semantically linked, so you can trace exactly how a data point relates to the rest of the market instead of looking it up in isolation.

Built for mechanical risk evaluation

Designed so portfolio and hedge risk can be assessed systematically against the underlying market structure — not reconstructed by hand from tribal knowledge each time.

Demystifies market structure

The complex relationships between physical exposure, hedges, benchmarks and time were previously the tribal know-how of experienced practitioners. We're putting that structure into our governed model.

Complementary, not competing

We sit alongside Exchanges, Price Reporting Agencies and ETRM platforms as the interpretive layer that explains what their data means — not another feed competing for the same shelf space.

Looking ahead

Born AI-Ready

As AI agents take on more of the work of analysing and acting on market data, they need to know not just what a price means but what they're permitted to do with it. We're building our Knowledge Engine so that usage rights and licensing policy guard rails travel with the data itself.

Ask it plainly
"What variation margin method do LME Aluminium Options use?"
searched the contract → traced its linked margin-method record
Net Liquidation Value (NLV) — the LME Primary Aluminium USD Option (contract AHD0, the standard "LME Aluminium Option") uses NLV as its variation margin method.

Built by people who've lived this problem

A founding team with direct, hands-on experience of the market MetalsData.AI serves.

Dominic Hall

Founder

A metals trader by background, Dominic co-founded and was CEO of Fastmarkets — the independent data vendor and metals price-reporting business acquired by Euromoney — before starting MetalsData.AI to build the reference-data infrastructure the market has never had.

He's now building the senior team alongside him, drawing on decades of combined experience across metals markets, market data and data businesses.

Register your interest

We're busy building all of this now. Tell us a little about yourself and how MetalsData.AI might help your business and we'll be in touch as things progress.

We'll only use this to get in touch about MetalsData.AI. No spam, no third parties.