AI Agent Operational Lift for Sibanye-Stillwater Reldan in Fairless Hills, Pennsylvania

Why AI matters at this scale

Sibanye-Stillwater Reldan, operating as Reldan, is a mid-market precious metals recycler based in Fairless Hills, Pennsylvania. With 201–500 employees and an estimated $120 million in annual revenue, the company refines gold, silver, platinum, and palladium from industrial and electronic scrap. At this size, Reldan faces the classic mid-market challenge: it must compete with larger, more automated refiners while lacking the vast R&D budgets of global players. AI offers a way to leapfrog traditional efficiency barriers without massive capital expenditure.

The AI opportunity in metals recycling

Precious metal refining is a data-rich environment. Every lot of scrap has a unique composition, and small adjustments in temperature, chemical fluxes, or processing time can significantly impact recovery rates. AI can ingest historical assay data, real-time sensor readings, and operational logs to prescribe optimal settings for each batch. This directly translates to higher yields and lower waste, with a potential margin uplift of 2–5%—a substantial gain in a thin-margin commodity business.

Three concrete AI use cases with ROI

1. Recovery yield optimization – By training a machine learning model on thousands of past refining runs, Reldan can predict the best process parameters for incoming scrap. Even a 1% improvement in gold recovery on a $100M throughput could add $1M to the bottom line annually, paying back any AI investment within months.

2. Predictive maintenance for furnaces – Unplanned downtime in a smelting furnace can cost $50,000–$100,000 per day in lost production. AI-based predictive maintenance using vibration and temperature sensors can reduce such events by 30–50%, delivering a clear ROI through increased uptime and extended equipment life.

3. Automated scrap sorting with computer vision – Manual sorting is slow and inconsistent. A vision system that classifies scrap by metal type and purity can speed up processing, reduce labor costs, and ensure a more homogeneous feedstock, further boosting downstream recovery.

Deployment risks specific to this size band

Mid-market companies often struggle with data silos and legacy systems. Reldan likely has operational data spread across spreadsheets, on-premise databases, and PLCs. Integrating these into a unified AI platform requires upfront IT investment and change management. Additionally, the workforce may be skeptical of AI-driven recommendations; a phased rollout with operator-in-the-loop validation is essential. Cybersecurity is another concern, as connecting industrial controls to the cloud expands the attack surface. Starting with a low-risk, high-return pilot (like yield optimization) and partnering with an experienced industrial AI vendor can mitigate these risks and build internal buy-in.