From 2008 to the first quarter of 2026, Eskom’s tariffs climbed fourfold in constant rand terms. The impact on the economy has been profound. From the mid-2010s a wave of closures hit energy-intensive producers. Other users found ways to cut their use of electricity and invested heavily in cheaper, more reliable renewable generation. For the past few years private investment in electricity infrastructure has thus exceeded Eskom’s. The scale of Eskom’s tariff hikes relative to the economy emerges from its growing share of GDP. Its sales of electricity in kilowatt hours dropped 15% from 2008 to early 2026. Yet, its revenues rose from 1.8% of GDP in 2008 to 4.7% in its 2024/25 financial year and are likely to reach 4.8% in 2025/26. In effect, Eskom now absorbs a rising share of national value added, at the cost of downstream producers and households. The most obvious effect has been on the aluminium, ferroalloys and steel refineries, for which electricity is often the most important cost. Many expanded or started production in the 1990s, when South Africa advertised some of the world’s cheapest electricity. Now, in contrast, electricity tariffs are about average for upper-middle-income economies. Moreover, South Africa’s unusual heavy dependence on coal increasingly exposes exporters to foreign carbon taxes.In 2025, the main metal refineries — in aluminium, ferrochrome, steel and manganese — consumed almost 20% less electricity than in 2008. The industry had experienced a series of plant closures, leading to a steady decline in the metals output. In every case, the owners put at least some blame on the rising cost of electricity. As a result, the share of chrome ore processed locally fell from about two thirds before 2008 to less than a seventh in 2026. For manganese, the decline was from a quarter to a tenth and for steel, from a third to a seventh. Output from the aluminium refineries, which use imported ore, dropped 10% in 2008-26, despite a preferential pricing agreement with Eskom. The ferrochrome crisis underscores the trend. The number of ferrochrome smelters in operation dropped from about 50 in 2008 to fewer than 25 in 2024 and just 11 in early 2026. From the late 2010s to 2025, South Africa’s ferrochrome production fell from more than 4-million tonnes to about 1.6-million tonnes, returning to levels last seen in the mid-1990s. The producers explicitly blamed soaring electricity costs, since energy is the largest single input into ferrochrome. In response, Eskom has offered to supply electricity to the smelters at a third of the average tariff, leading to the reopening of some smelters. More broadly, escalating Eskom tariffs have boosted the attractions of renewable energy, despite high initial investment costs. In response, private investment in electricity generation on and off the national grid has soared, cutting into Eskom’s sales. Before 2008, only about 10% of annual gross fixed capital formation in electricity infrastructure was private; for the past three years, it has averaged just more than 50%. As these figures largely exclude roof-top solar, they are understated. Eskom’s share in national electricity demand on and off grid has thus fallen from more than 90% before 2008 to less than 80% now. The Integrated Resource Plan anticipates a further fall to below 50% over the coming 15 years. For decades, South Africa relied on cheap electricity to promote production of crude metals. The economics of the electricity sector mean that the growth path can no longer be sustained. As with any fundamental disruption, it is tempting to try to turn back the clock rather than building on the innovative responses that have emerged in the rest of the economy. In particular, subsidising electricity for energy-intensive producers can only provide short-run relief. In the long run, it cannot substitute for strategies to ensure more fundamental and efficient adaptation to South Africa’s new energy realities.• Makgetla is a senior researcher with Trade & Industrial Policy Strategies.