Permutable expands AI commodity forecasting to 19 markets

Permutable has expanded its large language model-based commodity forecasting system from six markets to 19. The latest version recorded a 24.8% return in live evaluation for 2026 through 22 September.

The London-based fintech said the broader coverage spans energy, agriculture and metals. Cumulative performance since live tracking began in October 2024 reached 74.0% by 22 September 2026.

Permutable publishes market signals generated by a system that reads commodity news, company reporting, policy announcements and other public information. The model keeps a record of the sources behind each signal so users can review the material that informed a conclusion.

Permutable compared its 2026 live result with approximately 13.2% for the S&P 500 over the same period. The model also remained ahead of the S&P 500 during 2026 but trailed the rise in the Bloomberg Commodity Index, which Permutable presented as a reference benchmark rather than a like-for-like comparison.

Across the full live period from 3 October 2024 to 22 September 2026, the indexed cumulative level for the model reached 174.0, versus 141.5 for the Bloomberg Commodity Index and 136.2 for the S&P 500. The figures were based on 105 weekly marks.

Published 2026 data showed a maximum peak-to-trough decline of 5.3%, annualised volatility of 13.1% and positive weeks in 65.8% of observations. According to Permutable, energy, agriculture and metals made positive contributions.

How it works

Permutable said the system is designed to detect market narratives rather than rely on a conventional sentiment score. The model aims to distinguish between general attention and news that reflects market pressure, while preserving an audit trail of the source material behind each signal.

The expansion reflects a broader push by financial technology groups to apply generative AI and language models to investment research, particularly in markets where geopolitics, regulation, supply disruptions and weather can quickly move prices. Commodity markets are especially exposed to these cross-currents because shocks in oil and gas can affect industrial metals and farming costs, while sanctions and trade restrictions can alter shipping routes and buyer behaviour.

Permutable said its wider market coverage allows it to track when stress in one commodity begins to appear in another. That matters for investors and analysts trying to understand links across fuel, industrial inputs, food and precious metals rather than treating each market in isolation.

The company provides the data through an API for use in research and investment systems. The service includes live market signals, narrative drivers, source evidence and a record of how each signal was formed.

Performance detail

The release of market-by-market results is a notable part of the announcement because firms in the AI forecasting sector do not always disclose where a model has underperformed as well as where it has worked. Permutable said it has published detail on both stronger and weaker areas across the 19 markets.

Permutable also drew a distinction between its paper evaluation and an investable fund return. The results reflect live point-in-time signals rather than a traded product and exclude items such as transaction costs, roll costs, fees and execution slippage.

Permutable added that the modelled portfolio can take both positive and negative positions and uses leverage, unlike the conventional long-only indices used as reference points. That means headline comparisons with broad equity or commodity benchmarks should be treated with caution.

Wilson Chan, Founder and Chief Executive Officer of Permutable, said the wider scope marked a shift from proving the approach in a small group of markets to applying it across a much broader set of commodities.

Chan said: "Last year showed that our approach could identify meaningful changes in a focused group of commodity markets. This year is about breadth. We can now apply the same explainable intelligence across the major commodity complex, helping investors understand not only what is changing, but why markets may be moving together."

Michael Brisley, Chief Commercial Officer of Permutable, said: "Investors are looking for a coherent picture of markets that are increasingly connected by geopolitics, climate, regulation and supply chains. AI can help make those relationships visible as they develop, with enough evidence for investors to assess what they mean for their decisions."