Eskom's proposal to offer cryptocurrency miners a discounted price for surplus electricity has garnered support from the crypto sector. However, according to Stephen Boyke Sidley of partner Bridge Capital, the success of this initiative will depend on the size of the discount, the criteria for selecting participants, and the mechanisms for disconnecting miners during periods of network load.
The energy regulator Nersa is conducting consultations regarding Eskom's proposed two-year pilot tariff for clients engaged in cryptocurrency mining, as well as for other consumers with flexible loads. This tariff sets an indicative price of around 1.20 rand per kWh. The deadline for submitting comments is October 23rd at 16:00, and public hearings are scheduled for November 3rd; Nersa's decision is expected on December 7th.
Sidley, a practice professor at the University of Johannesburg who has written two books on Bitcoin and cryptocurrencies, noted that the idea itself is obvious: selling electricity at a discount when it is not needed by anyone. Nevertheless, he stressed that 'the devil is in the details.' He raised questions about the economic viability of the discount, the existence of other consumers with 'temporal flexibility' besides Bitcoin miners, and whether miners voluntarily disconnect during peak hours or if automatic telemetry is used. He agreed, however, that it is undoubtedly a good idea for Eskom provided the details are thoroughly worked out.
Christo De Wit, Country Manager for South Africa at the cryptocurrency exchange Luno, called the proposal 'innovative and forward-thinking,' noting that it 'directs surplus electricity towards productive use.' The Nersa consultation document, published last week, partially addresses these questions while leaving some open.
Regarding pricing, the indicative price of 1.20 rand per kWh applies only to clients in the highest voltage category in the southern region; rates in other transmission zones and voltage categories will be adjusted according to Megaflex, and customers receiving power below 500V are excluded.
Based on Megaflex
The pilot project is based on the Megaflex tariff, which Eskom uses for large clients, but most daytime hours on weekdays, which usually have a standard Megaflex rate, will become off-peak demand hours. The document states that operational windows will be approximately 12–16 hours per day. Since the full demand, capacity, and network charge schedule has not yet been published, the full cost of participation cannot yet be estimated.
The pilot is open to 'other qualified flexible customers,' and Nersa is determining whether cryptocurrency mining should be considered a 'separate client category' or if the tariff should be available to all clients meeting certain technical and operational requirements.
In April, Agnes Mlambo, then acting head of Eskom's distribution department, noted that producers were having difficulty utilizing surplus energy: 'It is not easy in a production environment to increase or decrease output.'
The issue of disconnection is not resolved in the document. Participants 'may be required to take part in demand response programs,' and Nersa is asking stakeholders whether mandatory demand response capability, minimum consumption reduction requirements, response time, telemetry, accounting, and penalties for non-compliance should be included in the pilot.
Eskom's argument is based on the existence of surplus. The company estimates that reserve capacity can range from 5 GW to 7 GW at certain times, driven by daytime solar generation and low nighttime demand, and that in 2028, more than 5 TWh might be reduced in the scenarios under consideration.
Reduction means requiring wind and solar power plants to cease generation, which incurs costs because power purchase agreements compensate producers for the energy they were ready to supply. Instructions for reducing solar generation have increased from about 100 per month at the beginning of this year to over 1000 per month, and the National Transmission Company of South Africa (NTCSA) reported in July that claims under review and settlement decreased from about 2 billion rand in mid-June to 1.5 billion rand.
Two cryptocurrency mining companies expressed interest, according to the document, with a 'combined initial demand of approximately 10 MW and potential expansion up to 500 MW.'
The Texas Precedent
Sidley pointed to an 'international precedent, especially in Texas.' Riot Platforms is one example. In the first quarter of 2026, the miner, whose shares are listed on Nasdaq, earned $21 million USD from electricity credits: $13.5 million was mainly derived from selling contracted electricity back to the grid instead of using it, and $7.5 million from demand response programs managed by the Electric Reliability Council of Texas (ERCOT) and its North Central counterpart MISO. After deducting these credits, Riot's electricity cost was 3 US cents per kWh.
This strategy is not new. In August 2023, Riot announced that it earned approximately $31.7 million in credits after 'reducing electricity consumption by more than 95% during peak demand periods' in the ERCOT grid.
A person involved in the Bitcoin mining sector in South Africa, who wished to remain anonymous and whose business could benefit from a cheaper tariff, asserted that miners offer exactly what the surplus grid needs. 'Eskom has a problem with excess capacity, and they need interruptible, intelligent, programmable load to stabilize the grid,' he stated. He added that ASIC-based mining installations, specialized chips used for Bitcoin mining, 'can be turned on and off at will, like boiling a kettle,' which, in his opinion, could help keep the grid frequency 'stable at 50 Hz.'
However, the pilot represents a tariff, not a contract for grid balancing services. The NTCSA system operator sets separate technical requirements for frequency reserves: instantaneous reserve must be fully available within 10 seconds. Loads can provide this through demand response, but the document leaves open what response times and penalties will apply to miners.
Furthermore, the proposed price is not low enough for a Bitcoin insider. 'No, you need to achieve a price below $0.05 for it to be economically viable at all. But it is a good start,' he said. At an exchange rate of about 16.36 rand to the dollar at the beginning of September 30th, $0.05 is about 82 cents. He did not specify if this is the full cost.
He also objected that rising tariffs force customers to opt out of the grid, citing 'complete demand collapse.' Eskom sales fell by 6.2% to 178 TWh in the 2026 financial year, although the company's profit more than doubled.
Caveats Remain
One potential operator is more optimistic. According to the document, they indicated that a price of around 1.30 rand per kWh with a daily window of 14–16 hours 'could be commercially viable and sustainable,' although the document does not specify if this is the full cost, and Nersa notes that such indications are not 'independent proof' of viability.
Nersa warns that, if not managed, cryptocurrency mining 'could create additional pressure on system capacity, increase emissions, and cause price distortions,' and preferential tariffs could create 'risks of unintended cross-subsidization or shifting costs to other customer categories.' Eskom states that the pilot should be revenue neutral in the 2027 financial year.
There is a precedent for protecting other customers. When Nersa approved a preferential tariff of 62 cents per kWh in May for Samancor Chrome ferrochrome smelters for five years and Glencore-Merafe for three, it stipulated that any revenue deficit must be isolated within Eskom and not compensated for by standard tariff customers.
} , 3. {