SA is sitting on the hardware of the AI economy

Share current article via Email Share current article via Facebook Share current article via Twitter Share current article via LinkedIn While South Africa supplies a disproportionate share of the physical inputs required for the AI age, the writer says it remains locked into the lowest-margin segment of the value chain. South Africa supplies the physical building blocks of the AI age, yet captures only a fraction of the value that sits downstream of the mine. The country is endowed with far more critical minerals than currently enter global supply chains. It holds the overwhelming majority of the world's platinum group metal (PGM) reserves and remains one of the largest producers of manganese. These are not niche industrial commodities. They are foundational inputs for the infrastructure that powers AI: catalysts, sensors, specialised alloys, high-performance electronics and the energy systems that keep data centres running. According to Valterra Platinum (formerly Anglo American Platinum), conservative estimates put current PGM demand linked to AI-related applications at 200,000-300,000 ounces a year, with the potential to grow four- or five-fold in the next five to eight years. Manganese demand for battery chemistries is rising in parallel as the world builds storage capacity for the energy transition and the electricity-hungry AI economy. Most manganese still leaves the country as ore. Local smelting capacity has contracted sharply due to unreliable, expensive power. While some PGM producers maintain sophisticated refining operations on the highveld, the broader pattern remains one of exporting bulk commodities and importing finished high-value materials and equipment. The result is that the country supplying a disproportionate share of the physical inputs stays locked into the lowest-margin segment of the value chain. This is not an inevitable outcome; it is a positioning choice. The global scramble for secure, diversified supplies of critical minerals creates a rare opening for producers that can demonstrate more than geological endowment. Amid geopolitical tensions across multiple regions, South Africa is well placed as an alternative supplier to the Americas, Europe, Africa and Asia — provided transport, logistics, energy and skills keep pace. Investors and off-takers look for jurisdictions and companies that can deliver reliable, ESG-compliant midstream and downstream capability. South African miners are closer to that opportunity than most of their peers on the continent, provided they treat the current moment as a transformation challenge rather than another policy debate. The following practical levers stand out: Energy reliability must be reframed as an operating-model issue, not merely a power-purchase agreement. Private generation, wheeled renewables and sophisticated energy-management systems are already being deployed at scale by leading operators, alongside government efforts to advance the National Transmission Company and energy trading markets. The next step is to integrate energy data with production systems so that power availability becomes a controllable variable rather than a constraint. Mines that can guarantee stable, lower-carbon electricity for processing plants will be far more bankable as partners for battery materials, speciality alloys and other AI-adjacent products. Digital operating models and data foundations are the missing link between resource and value. Beneficiation projects fail not only because of capital or policy but also because of variability, yield losses and opaque cost structures. AI-enabled process control, predictive maintenance and real-time geometallurgical modelling can compress those risks. The same platforms that improve safety and productivity in the pit can make local processing commercially viable. Companies that treat data quality and digital capability as core assets will find it easier to attract technology partners and development capital for downstream facilities. Skills and partnerships must be treated as strategic rather than residual. Capturing more of the AI value chain requires metallurgists, process engineers, data scientists and commercial teams who understand the resource and the end-markets. Managed services and platform approaches can accelerate capability while allowing mining houses to keep scarce leadership attention on the core business. Selective joint ventures with technology providers, battery manufacturers and specialised processors can also derisk the step into higher-value products without requiring every company to reinvent the entire chain. None of this requires waiting for a perfect policy environment. The most successful moves will be those that start with what individual operators can control: the reliability of their energy systems, the quality of their operational data, the depth of their technical partnerships, and the clarity of their value-chain ambition. Policy can accelerate or hinder, but it cannot substitute for commercial readiness. • Botha is mining & metals leader at EY Africa. Business Day