Surge-Grade Commodity Operations: Real-Time Liquidity, Margin, and COMEX–London Basis Control

Image
Chris McManaman

Opening Insight: Precious‑Metals Surges, Cross‑Venue Linkages, and Intraday Operating Model

Record sessions in precious metals—across COMEX, MCX, OTC, and ETFs—have tightened cross‑venue linkages, raised positioning, and amplified COMEX–London basis whipsaws. What changed is the tempo: intraday is now the operating unit. VaR shocks, margin cascades, and patchy liquidity collide with manual workflows and batch‑era systems, leaking margin, distorting hedges, and inviting audit risk. This isn’t an outlier; it’s the new equilibrium.

This post lays out how to operate when the screens light up. First, the cost of inaction and the benefits of fixing it. Then a surge‑grade operating model: a real‑time, event‑driven fabric; autonomous decisioning for margin and collateral; basis‑aware ETRM modernization across OTC and listed flows; ML‑driven forecasting; and a control plane built on rules‑as‑software, lineage, and model governance. We translate the model into architecture, KPIs and guardrails, a sequenced roadmap (including a four‑week Volume Surge Readiness Assessment), and explicit trade‑offs—events versus batches, sidecar services versus core ETRM extension, and automation with human oversight. We close with FAQs and an implementation playbook to convert volatility into liquidity protection and basis capture. We now turn to Context and Analysis, which details the market structure shifts, flows, and basis behavior underpinning this surge regime.

Risks of Ignoring Surges in Precious‑Metals Trading

Ignoring surge‑grade operating challenges turns record precious‑metals activity into avoidable losses and control failures.

Left unaddressed, these stresses stack into margin leakage, distorted P&L, audit flags, and compounding fragility.

that becomes a persistent competitive handicap.

Benefits of Solving Surges

Surge‑Grade Operating Model

A surge‑grade operating model unifies real‑time telemetry, automated decisioning, and firm‑wide controls. On record‑volume days—when global trading ran ~US$410bn/day versus a ~US$361bn 2025 average, ETFs sat at US$559bn on US$89bn of inflows, and OTC rotated 16% m/m—it protects liquidity, contains basis drift, and acts intraday. It also catches issues like the ~12% VaR under‑call before they become P&L.

Surge-Grade Architecture and Roadmap

Record sessions across COMEX, MCX, and ETFs turned

When flows rotate fast and liquidity gaps, end-of-day processes miss the window. On a record day, our own surge model under-called VaR by ~12% around lunch. The operating model must therefore be surge-grade : real-time, automated where routine, and tightly controlled so liquidity is protected and basis edges are captured without audit surprises.

Event-Driven Architecture for Intraday Margin, Credit, and Basis

Control and Governance for Explainable Automation

ETRM Integration Across Venues and Counterparties

Data Models and Streaming Signals

KPIs, Guardrails, and Auditability

Roadmap and Sequenced Delivery

Practical Trade-offs to Manage

Human/Org actions and roles

Ownership

Executive FAQs on Surge Days

Which signals should we monitor to anticipate surge days?

Watch synchronized ETF inflows and AUM highs—gold‑backed ETFs reached US$559bn on US$89bn of annual inflows—alongside volume jumps like December’s ~US$410bn/day versus the 2025 average of ~US$361bn/day. Track positioning: COMEX net longs at 683t and money managers at 395t reinforce liquidity depth and trend conviction. Monitor OTC–listed rotation (OTC up 16% m/m while exchange‑traded derivatives fell 16% m/m) and the COMEX–London basis; on 12 Dec 2025 the front month ran roughly US$10.4/oz over London.

How do we keep margin and liquidity under control when volumes spike?

Expect intraday VaR jumps and margin‑call cascades; end‑of‑day risk leaves treasury and credit blind. Stream prices, positions, margin, and credit usage as events, and scale compute with cloud elasticity. Use autonomous decisioning to triage calls and propose collateral substitutions, codifying collateral and credit rules as software. Modernize ETRM for intraday revaluation and multi‑venue, basis‑aware hedging, and convene intraday risk councils to act in minutes.

What actually breaks if we don’t adapt?

Manual confirmations, allocations, and cash movements don’t scale during record sessions, and suitability controls lag retail‑like flows. You’ll leak margin, misprice credit, and trap cash as basis/location dislocations distort hedges and settlements. Expect higher latency and error rates, audit flags, operational bottlenecks, and counterparty exposure—ending in competitive disadvantage.

Operate for the Surge Regime

Record single‑day surges in gold futures, reinforced by ETF assets at US$559bn on US$89bn inflows and December’s ~US$410bn/day turnover versus a ~US$361bn/day 2025 average, exposed the core risk: intraday VaR and margin shocks now collide with basis moves and cross‑venue linkages while workflow and latency limits turn into P&L. This is a recurring regime, not an anomaly—flow is rotating between OTC, funds, and listed markets, retail participation is reshaping books, and portfolio plumbing

Across credit, collateral, and settlements strains under pressure. The leadership task is to normalize intraday decision cycles: fuse real-time telemetry with automated, controlled actions so trading, treasury, and risk move in one rhythm. A surge-grade operating model protects liquidity, reduces trapped cash, and captures basis edges—durable advantages when the screens light up.

Implement Surge‑Grade Operations

Liquidity surges are now routine, turning margin, basis, and risk latency into intraday problems. Arcelian connects market structure with operating reality and helps leaders stand up a surge-grade operating model—event-driven, autonomous decisioning, and strong controls—to protect liquidity and capture basis edges. We turn record-day lessons into resilient, executable workflows.

Next step: commission a four-week Volume Surge Readiness Assessment.

Agentic AI in Commodity Trading: Integration Choices and Operating Trade-offs

Record-volume days in precious metals expose the limits of batch-era operating models. An effective modernization strategy starts with an event-driven fabric that unifies real-time telemetry from trading venues (COMEX, MCX, OTC, ETFs), treasury cash ladders, intraday VaR, and credit exposures—so agents can triage margin shocks and execute collateral substitutions within seconds.

The core design choice is whether to extend the ETRM architecture for streaming revaluation and basis-aware positions, or to deploy sidecar services that subscribe to event buses and write back decisions via APIs. Selection criteria should include latency budgets for intraday revaluation, deterministic sequencing of events (pricing → exposure → margin → funding), lineage requirements for model outputs, and the tolerance for coupling core books-and-records with autonomous decisioning.

A pragmatic integration roadmap sequences capability while containing risk. Phase 0 establishes streaming market data, order/position telemetry, and golden-source reference data with lineage. Phase 1 introduces agentic services for margin call triage and collateral optimization with rule-as-software guardrails, explicit entitlements, and human-in-the-loop thresholds. Phase 2 expands to basis-aware hedging across COMEX–London with cross-venue netting and liquidity-aware routing. Phase 3 automates exception-based STP in credit, confirmations, and settlements, with policy controls,

auditability, and kill-switches governed by Model Risk. Cloud elasticity is used for intraday bursts, with cost caps and autoscaling policies governed by KPIs rather than static capacity. This advances the post’s central thesis that event-driven, tightly governed automation—not isolated models—is the lever for liquidity protection and basis capture under volatility.

Measure progress through operating KPIs:

Trade-offs remain: build vs buy for agent frameworks, sidecar agility vs core ETRM complexity, and lower latency vs eventual consistency. Governance and controls must scale with autonomy.

Frequently Asked Questions

What signals should we watch to spot a surge day early?

Track synchronized ETF inflows and AUM highs (gold‑backed ETFs reached about US$559bn on US$89bn annual inflows), volume spikes (December averaged ~US$410bn/day vs a ~US$361bn/day 2025 average), and positioning (COMEX net longs ~683t; money managers ~395t). Watch OTC–listed rotation (OTC up ~16% m/m while exchange‑traded derivatives fell ~16% m/m) and the COMEX–London basis; on 12 Dec 2025 the front month was roughly US$10.4/oz over London. When these move together, cross‑venue linkages tighten and intraday basis moves accelerate.

How can we keep margin and liquidity under control when intraday volumes spike?

Replace batch updates with an event‑driven fabric that streams prices, positions, margin, and credit in real time; use autonomous agents to triage margin calls and propose collateral substitutions with exceptions escalated to humans; modernize ETRM for intraday revaluation and multi‑venue, basis‑aware hedging; codify collateral and credit rules as software; and use cloud elasticity for burst risk runs. Relying on end‑of‑day recalcs can under‑call intraday VaR (we saw ~12% around lunch on a record day) and drain liquidity.

Where should we start to build a surge‑grade operating model in our ETRM/CTRM stack?

Begin with a four‑week Volume Surge Readiness Assessment to quantify gaps. Design an event‑driven operating fabric and API/webhook interfaces, refactor ETRM for intraday revaluation and COMEX–London basis handling, and deploy agents for margin triage and collateral substitutions. Codify rules‑as‑software with model governance, instrument data quality and lineage, enable ML‑driven liquidity/margin forecasting, and scale compute with cloud elasticity.

Trend Watch

Agentic AI is shifting from advisory to execution in commodity trading: autonomous decisioning layered

on an event-driven operating fabric is becoming table stakes for surge-grade operations. Record commodity trading volumes in COMEX gold futures and MCX gold futures, reinforced by persistent gold ETF inflows, are tightening cross-venue linkages and amplifying COMEX London basis whipsaws. On those days, intraday margin calls cluster, credit haircuts ratchet, and treasury liquidity gets pinned unless software agents act within seconds.

What leading desks are deploying now:

The outcome: faster capture of basis edges, lower P&L variance, and preserved optionality when screens light up. Teams that pair agentic AI with ETRM modernization and real-time telemetry will manage surge days as a routine—and convert COMEX London basis volatility and OTC rotation into opportunity.

Start by pressure-testing latency and data gaps with a Volume Surge Readiness Assessment , then sequence agent rollouts to the highest-value choke points in margin, collateral, and hedge routing.

Closing Insight

Surge conditions are now the baseline, and advantage will belong to desks that operate for the intraday regime—not explain it after the close. Pair agentic AI with an event‑driven fabric and basis‑aware ETRM modernization so price→exposure→margin→funding executes deterministically in seconds, with model governance, lineage, and rules‑as‑software keeping risk management explainable. Codify collateral and credit policies, stream real‑time telemetry, and use cloud elasticity to scale VaR, liquidity forecasting, and treasury exception handling; the payoff is liquidity preservation, lower P&L variance, and systematic capture of COMEX–London basis edges. Start with a Volume Surge Readiness Assessment to surface latency and data gaps, then sequence agents at the highest‑value choke points—margin triage, collateral substitutions, hedge routing—so resilience and optionality compound every time volatility spikes.

Partner with Arcelian

Record-volume sessions are now the operating baseline; the advantage goes to firms that can align price→exposure→margin→funding in seconds with explainable controls. Arcelian partners with CIO, COO, and Treasury leaders to stand up surge‑grade operations—event‑driven

fabrics, basis‑aware ETRM modernization (COMEX–London), and agentic decisioning for margin and collateral—so liquidity is protected, P&L variance compresses, and auditability strengthens.

Connect with our team to evaluate your surge readiness and design a sequenced roadmap—from stress tests and data lineage to autonomous playbooks—that turns intraday VaR shocks, basis moves, and OTC–listed rotation into repeatable performance.

Subscribe to The Arcelian Brief

⚙️ Stay ahead of energy market shifts, trading intelligence, and the latest on AI-driven modernization.

Chris McManaman is the Managing Director of Arcelian, where she leads enterprise transformation initiatives that merge advanced analytics, agentic AI, and operational modernization across the global energy and commodities sectors. With over 25 years of experience in consulting and software strategy, Chris has built a reputation for turning complex systems into measurable business outcomes. Her career spans leadership roles in product strategy, digital transformation, and supply chain transparency, with deep expertise in process automation, data governance, and emerging technologies including AI, blockchain, and IoT. At Arcelian, she drives a mission to help energy and industrial companies bridge the gap between innovation and execution—delivering solutions that are technically robust, operationally grounded, and built for scale.