How can South Africa harvest the compounding benefits of artificial intelligence without deepening our water and energy crises? The solution is a strategic pivot toward sovereign, circular, and green-data industrialisation.
South Africa’s industrial ambitions are standing at a historic crossroads. The Department of Trade, Industry and Competition’s (DTIC) Industrial Development Strategy (IDS) 2026 outlines a bold roadmap to counter deindustrialisation, reverse sluggish growth, and modernise the nation’s productive sectors. At the heart of this vision is a rapid transition into the digital age, driven by the global expansion of the Fourth Industrial Revolution (4IR) and Advanced Digital Production (ADP) technologies like Artificial Intelligence (AI).
Yet, as the physical infrastructure of AI – primarily massive data centres – begins to scale across the country, and the continent, a stark developmental tension emerges. AI is not an ethereal cloud; it is a resource-intensive physical industry. High-performance data centres consume enormous volumes of electricity for computational processing and require significant amounts of water for cooling. In a country grappling with a long-standing energy distribution crisis and an escalating, climate-induced water security crisis, South Africa’s digital industrialisation must navigate a narrow bottleneck: how to harvest the compounding benefits of AI without draining the very resources that sustain its citizens.
To resolve this paradox, South Africa must move away from viewing AI and data centres as mere resource burdens. Instead, through strategic coordination, sovereign capability building, and circular resource management, the country can leverage this digital wave to accelerate its green transition and anchor its industrial future.
The industrialisation imperative
South Africa’s current position in the global digital economy is one of consumerism rather than creation. According to the United Nations Industrial Development Organisation (UNIDO), South Africa is categorised as a “Follower in Use” of advanced digital technologies, functioning primarily as a “User” rather than a “Producer” of ADP systems. Historically, a mere ten economies have monopolised 90% of global patents and 70% of exports directly associated with these advanced systems.
For South Africa to truly re-industrialise, this status quo is unsustainable. As Thabo Kekana, Deputy Director-General of Energy Programmes and Projects at the Department of Electricity and Energy, warned at Enlit Africa 2026:
“Digitalisation must become a platform for Africa’s industrial capability, not a new technological dependency. South Africa must build a sovereign digital capability through its engineers, software developers, data governance, cyber security standards and innovation ecosystems.”
To transition from a digital follower to a sovereign producer, the country’s National System of Innovation (NSI) must pivot its skills base. This includes supporting postgraduate studies in computer, information, and advanced digital sciences, alongside the urgent establishment of a new Advanced Engineering and Digital Technologies Science Institute to catalyse this nascent industry. Data centres and AI are the core infrastructure of this sovereign capability; they are the “factories” of the 21st century. However, building these digital factories requires confronting the realities of South Africa’s resource constraints.
AI as the nervous system of energy
The most immediate bottleneck for data centres is electricity. South Africa’s power sector is undergoing its most significant structural reform in decades, shifting from a centrally managed utility model to a more dynamic, market-based system. In this evolving landscape, data centres represent a dual reality: they are highly demanding energy consumers, but they are also potent catalysts for grid modernisation.
Thabo Kekana appropriately describes digitalisation as the “nervous system” of the modern electricity grid [135, 305]. Advanced systems, digital twins, and AI-driven predictive maintenance are critical to improving demand forecasting, anticipating transformer faults, and integrating distributed renewable energy. However, Kekana delivered an essential reality check:
“But let us be clear: AI cannot substitute generation capacity, the algorithm cannot replace transmission lines, dashboards cannot repair our transformers. Technology compounds impact when joined to infrastructure, skills and institutions.”
The real power of AI lies in its ability to manage how and when electricity is consumed, rather than just how it is generated.
Case in point: Smart demand management
A powerful use case of this principle is the partnership between South African energy tech firm Plentify and residential developer Balwin Properties. Residential water heating is a massive, often overlooked driver of grid pressure, accounting for up to 40% of electricity consumption in residential developments.
By installing over 7,500 AI-driven hotbot geyser controllers, the collaboration optimised water heating based on solar availability, lower tariffs, and resident usage patterns. The results of this coordinated demand management are striking:
-
A 46% reduction in peak electricity usage.
-
A 36% reduction in short-term demand spikes.
-
A 79% increase in solar energy utilised for water heating.
-
Over 1,458 tonnes of CO2 emissions avoided.
This proves that AI can actively relieve pressure on a struggling grid. Furthermore, rather than acting as a net drain, large-scale, properly integrated data centres can serve as anchor off-takers that accelerate private sector investments in new utility-scale renewable generation, transmission expansion, and battery storage systems. Under the amended Electricity Regulation Act, data centres can leverage wheeling frameworks and green power products, effectively funding the build-out of the very grid they rely upon.
Water security and the circular economy
While energy dominates national discourse, water scarcity is a more absolute physical limit. South Africa is a chronically water-scarce nation. In the NSI’s strategic planning, water security is recognised as a critical societal mission under the “Climate Crisis/Change” grand challenge, which aims to pioneer integrated solutions for water security.
Because data centres require millions of litres of water daily to maintain optimal operating temperatures, locating them in water-stressed municipal areas risks direct competition with domestic and agricultural needs.
How can South Africa resolve this hydrological conflict? The blueprint may lie in the planning of another nascent resource-intensive sector: the Green Hydrogen economy.
The desalination blueprint
In his keynote address on South Africa’s green hydrogen economy, Mashopa Mabuya (presidency lead for green hydrogen) addressed the challenge of “scarce electrons and scarce water”. To prevent green hydrogen production from competing with municipal potable water, South Africa’s major gigawatt-scale projects are planned for coastal locations, leveraging the country’s extensive coastline to access water through desalination.
While desalination is energy-intensive, it decouples industrial growth from municipal water basins. However, Mabuya notes that this is not a consequence-free solution: projects must undergo rigorous environmental assessments to manage the density and ecological impact of the brine released back into marine environments.
For South Africa’s AI ambitions, a similar spatial and technological strategy is required. Establishing “Digital Special Economic Zones” along coastal regions – such as the Western Cape or Eastern Cape – would allow data centres to couple with seawater desalination networks, using sea-water cooling loops or desalinated water, thereby shielding inland municipal reservoirs.
The NCPC framework
For inland data centres where desalination is impractical, industrial water usage must transition from a “take, make, and dispose” extractive model to a circular, closed-loop system. The circular economy is designed to restore and regenerate resources, maintaining them at their highest level of value for as long as possible.
South Africa already possesses the institutional framework to drive this transition.
The National Cleaner Production Centre (NCPC), hosted at the CSIR and fully funded by the DTIC in partnership with UNIDO, assists industries in adopting Resource-Efficient and Cleaner Production (RECP) methodologies. By applying RECP principles to digital infrastructure, data centres can be designed to:
-
Narrow Water Flows: Implement advanced liquidless “free-air” cooling systems or high-efficiency closed-loop chillers that recycle 99% of their cooling water.
-
Create Resource Loops: Partner with local municipalities to utilise treated industrial wastewater (reclaimed water) for cooling rather than drinking-grade potable water.
-
Recover Waste Heat: Redirect the low-grade thermal waste heat generated by data centres to nearby agricultural or industrial processes (e.g., greenhouse warming or manufacturing pre-heating), embodying true circular industrial symbiosis.
The path forward
South Africa cannot afford to let its digital industrialisation be choked by a lack of coordinated planning. Historically, the National System of Innovation has suffered from fragmentation and siloed operations across government departments.
To build a sustainable AI economy, a unified, cross-governmental approach is vital. The DTIC, the Department of Science and Innovation (DSI), and the Department of Electricity and Energy must work in lockstep to align industrial incentives with resource realities:
-
Targeted Incentives: Pivot the DTIC’s manufacturing and industrial incentives, such as the Manufacturing Competitive Enhancement Programme (MCEP) or Special Economic Zone (SEZ) benefits, to reward data centres that construct their own renewable energy supply and implement 100% closed-loop circular water systems.
-
Infrastructure Fast-Tracking: Utilise the Infrastructure Development Act to register resource-efficient, co-located digital infrastructure projects as Strategic Integrated Projects (SIPs), drastically reducing bureaucratic turnaround times for environmental, water-use, and grid-connection approvals.
-
Sovereign Skills Development: Fund dedicated research chairs and collaborative platforms at universities of technology to train local engineers in green-data infrastructure, ensuring the country produces the technical competencies required to manage this specialized transition.
South Africa’s digital transition is not an optional luxury; it is the cornerstone of its future industrial competitiveness. The high electricity and water usage of AI industrialisation represents a formidable hydrological and energetic boundary, but boundaries are also the birthplace of innovation.
By demanding strict circular economy compliance, utilising coastal desalination blueprints, and using AI itself to balance the national grid, South Africa can transform a potential resource burden into a powerful engine for green re-industrialisation. Energy security and water security are not accidental; they are designed. Through institutional coherence and bold strategic intent, South Africa can design an AI landscape that is both technologically sovereign and ecologically sustainable.
