Structural Shifts: Commodities

Over the past five years, the commodities hedge fund landscape has undergone one of its most profound structural shifts in decades. Multi-manager platforms — once built around independent, siloed pods — are now moving toward shared analytical infrastructure integrating data, modelling, and research across desks.
This article explores three case studies: Millennium, DE Shaw, and Citadel. Each represents a different stage in the move toward centralised intelligence, where information, not just capital, becomes the foundation of competitive advantage.

Case Study 1: Millennium

Millennium’s commodities business is both the fastest growing and one of the most structurally complex within the firm. It also provides the clearest example of how a traditionally pod-based platform is evolving toward a centralised analytical model.

The process began when Millennium separated its commodities unit from the wider Fixed Income group in 2023, appointing Anthony Dewell (formerly of Goldman Sachs) as Head of Commodities Trading. This was the start of a multi-legged effort to bridge the gap between MLP and its peers and to open the path to a broader commodities effort, which could see the firm venture into physical markets for the first time. Under Dewell, the firm consolidated 31 trading teams into 14 “Super Pods”, a decisive move aimed at driving performance, reducing duplication, and improving collaboration. This restructuring marked Millennium’s fourth major attempt at building a sustainable commodities business — following earlier efforts that faltered due to fragmented analytical infrastructure.

Until recently, all research and modelling within commodities were produced at the pod level, with associated costs carried by individual PMs. When considering a shift to a more centralised model, senior management was initially hesitant, concerned that a unified analytics function might disrupt the culture and autonomy of the fixed income business, which is one of the main drivers behind the commodities team’s separation in the first place. However, the strategic calculus shifted in mid-2024, when Millennium began formally exploring a centralised analytics division under the leadership of Rossen Roussev (later succeeded by Stephen Waton, Head of Commodities Technology (formerly Citadel’s European Head of Commodities Technology) after Roussev’s departure to DRW in January 2025)

The new structure is built around three verticals:

• Quantitative Modelling, led by Abhishek Bagri,
• COT (Commitment of Traders) Analysis, under Chunli Guo,
• Data Onboarding and Infrastructure, managing ingestion pipelines and tool integration.

A cornerstone of Millennium’s new approach has been the expansion of its India-based analytics footprint, now one of the largest and most sophisticated among hedge funds. The quant team in India, led by Bagri, currently consists of five analysts, including Rohit Sarkar and Vignesh Natarajan Ganesh (both ex-Goldman Sachs).
Sarkar has led the build-out of Millennium’s US and European Power & Gas platform, while Ganesh brings experience modelling US power and gas markets across ISOs and researching renewables’ impact on supply–demand dynamics. Also on the team are Sanidhia Maheshwari (ex-Morgan Stanley) and Aditya Gupta, a Berkeley master’s graduate currently interning after previous experience at Goldman Sachs.

The group’s mandate is to support newly onboarded PMs, enabling them to become operational more quickly by providing pre-built models, data pipelines, and shared infrastructure — effectively shortening the time to deployment for new strategies.

This move toward centralisation marks a decisive cultural shift at Millennium under Anthony’s leadership. While its pod-centric model remains intact elsewhere in the firm, the commodities division is becoming a test bed for a more unified research framework. What remains to be seen is whether Millennium will continue to support embedded pod-level analysts in parallel with the centralised buildout or if this signals a longer-term shift toward a fully unified analytics infrastructure.

Case Study 2: DE Shaw

DE Shaw’s analytics platform has rivalled Citadel’s and is widely considered as best in class across the entire commodity trading landscape. DE Shaw’s main commodity fund, the Plasma Fund (est. 2011, co-founded by Ron Ozer), is renowned for its methodical approach to fundamental analysis. The fund is the primary investment vehicle for fundamental energy markets for DE Shaw, with a $3.5bn AUM and a mandate to trade directional and relative value strategies across energy, but with a focus on power and gas markets globally.

DE Shaw’s edge in energy markets is underpinned by its centralised twelve-person research team based in Hyderabad, India, led by Mayank Gupta, who founded the group in 2008. This team provides daily analytical support to trading desks in London and New York, modelling global oil, gas, and power balances. Gupta’s core expertise lies in quantamental research and statistical modelling, particularly in forecasting U.S. and global crude oil and refined product balances. However, as the first hire for DE Shaw’s Plasma Fund in Hyderabad, Gupta has also played a key role in building and mentoring the team, taking personal responsibility for training junior analysts recruited directly from the university.

DE Shaw’s hiring model within commodities reflects its approach of recruiting top-tier talent through internships and campus programs at India’s leading institutions. Most members of the Hyderabad team are graduates of the Postgraduate Diploma in Business Analytics (PGDBA) — a joint programme run by the Indian Institutes of Technology (IITs) with an acceptance rate of roughly 0.5%.
Analysts are placed into three specialist verticals:

1. Crude & Products,
2. EMEA Power & Gas,
3. U.S. Power & Gas.

For example, Piyush Raj Gupta, a senior analyst in U.S. Energy, focuses on LNG infrastructure, forecasting exports by modelling maintenance schedules, unplanned outages, and long-term capacity expansion. This depth of specialisation — coupled with coordination between the Hyderabad analytics hub and front-office trading desks — has given DE Shaw a clear competitive edge. Ideas can be tested, refined, and implemented rapidly, maintaining analytical consistency across continents.

However, the model is not without challenges: over the last three years, DE Shaw has struggled with talent retention, with senior analysts such as Naveen Kulkarni and Abhishek Bagri joining Millennium in Bangalore, and Adit Chopra relocating to Citadel in Calgary.

Case Study 3: Citadel

Citadel remains the benchmark for scale and integration in commodities investing.
Over the past four years, its commodities division has generated more than $15 billion in profits, underpinned by a deep, data-driven understanding of physical and derivative markets.

Citadel operates across nearly all major commodities, excluding base metals, which it famously exited and vowed to never re-enter following consecutive years of disappointing performance. Citadel’s investment philosophy is grounded in a deep, data-driven understanding of commodity fundamentals, which allows them to express views with high conviction underpinned by rigorous modelling and scenario analysis, centrally coordinated by Greg White, the Head of Commodities Research. He is supported by a network of over 100 engineers and data scientists embedded directly within trading teams. These engineers are treated as product specialists, contributing to strategy design, model development, and execution algorithms.

The firm’s commodities platform is distinguished by its tight integration of engineering and trading. Analysts and data scientists build production-grade Python code that feeds into a shared research environment accessible across teams, ensuring scalability and data consistency. This collaboration enables real-time analytics, scenario testing, and infrastructure maintenance on a global scale.

Citadel’s organisational model is deliberately collaborative — unlike many multi-strategy peers that operate in silos. Portfolio Managers contribute to a shared house view and actively engage with teams across products and regions, while Fundamental Analysts are aligned to senior PMs and focus on detailed supply and demand modelling, market dislocations, and trade idea generation that benefit the entire platform.

An example from within Citadel’s oil team illustrates this approach:
A former Crude Oil Analyst was responsible for pipeline and tanker flow models, Canadian WCS pricing forecasts, and refinery maintenance projections, all built using facility-level production data. This data was integrated into global refinery and logistics models, allowing for precise scenario testing and cross-team execution.

The result is a research ecosystem where every dataset, model, and codebase contributes to a unified analytical infrastructure — the foundation of Citadel’s scale and consistency.

Conclusion

While Millennium, DE Shaw, and Citadel each operate with distinct cultures and investment philosophies, they converge on one shared principle: analytics are no longer a support function — they are the infrastructure of performance. Millennium exemplifies incremental centralisation, DE Shaw shows the power of quant-engineering integration, and Citadel demonstrates how scale and coordination can produce sustained alpha. The race to build centralised intelligence systems within modern commodities trading increasingly determines who leads in this space.