From Soil to Silicon: Does AI Redefine South Africa’s Land Question?

Introduction

Land reform in South Africa remains one of the most enduring and complex socio-economic policy challenges of the post-apartheid era. The historical roots of the country’s land question trace back to the 1913 Natives Land Act, which legally restricted Black South Africans to as little as 7% of the country’s land, later expanding to 13% by 1936 a drastic distortion of land ownership that maintained centuries of exclusion from productive assets and economic opportunity[i]. This systemic deprivation laid the foundation for deep-rooted inequality that persists nearly 30 years into democracy, with white commercial farmers still owning a disproportionate share of productive land and redistribution targets unmet[ii].

In the conventional economic framework, land is a fundamental factor of production alongside labour and capital. Access to land enables agricultural livelihoods, supports rural economies, underpins housing and infrastructure development, and serves as a store of wealth and collateral. For South Africa, equitable access to land has been framed not only as a matter of justice but as a mechanism for poverty alleviation, employment creation, and structural transformation, objectives rooted in the National Development Plan and successive policy instruments.

Yet we are now living in a period of rapid technological change. Artificial Intelligence (AI), once confined to research laboratories, is reshaping economic structures at unprecedented speed. Globally, enterprise AI markets are expanding sharply. For example, South Africa’s own enterprise AI sector generated an estimated USD 414.7 million in 2024 and is forecast to nearly septuple to USD 2.88 billion by 2030, growing at an annualized rate exceeding 39%[iii]. At the same time, generative AI technologies and machine learning tools are being applied across sectors such as healthcare, finance, manufacturing and agriculture, enhancing productivity, reducing transaction costs, and enabling new digital business models that rely more on data and compute than on physical land[iv].

These developments raise a pressing question for South Africa’s policy discourse: Does the rapid adoption and integration of AI materially challenge the traditional economic case for land reform? If production increasingly leverages intangible digital assets, algorithms, and automation, does land’s role as a determinant of economic opportunity diminish, or does it remain essential, especially in contexts where land rights are intimately tied to identity, livelihoods, and socio-economic inclusion?

This article explores this question by examining the evolving role of land as a production input in the age of AI, the implications of AI-driven economic transformation for rural development and inequality, and how land reform policy might need to adapt to ensure inclusive and sustainable growth in South Africa’s Fourth Industrial Revolution.

The Traditional Role of Land in Economic Development

Historically, land has occupied a central position in South Africa’s development trajectory as both a productive asset and a store of wealth. Productively, land underpins key sectors of the economy. Commercial agriculture, which despite contributing only around 2–3% of GDP in 2025, remains critical for rural employment and food security[v]. Mining, which is inherently land-dependent and accounts for roughly 7–8% of GDP and over 50% of export earnings[vi]. Additionally urbanisation and industrialisation which anchors manufacturing, logistics, housing, and municipal revenue through rates and services[vii] are also anchored by land. Beyond production, land functions as a primary vehicle for wealth accumulation and intergenerational transfer, particularly through property ownership, collateralisation, and spatial proximity to economic opportunity. In South Africa, the systematic dispossession of land through colonial conquest and apartheid legislation translated directly into persistent poverty and inequality. By 1994, Black South Africans, comprising nearly 80% of the population, owned a negligible share of productive land, entrenching racialised asset poverty that remains visible today. Land reform was therefore conceived not merely as redistribution of hectares, but as a mechanism for economic empowerment, inclusion and redress, aimed at restoring access to productive assets, correcting spatial distortions and enabling historically excluded households to participate meaningfully in economic growth. This traditional framing places land at the heart of development policy, both as an economic input and as a structural determinant of inequality.

Artificial intelligence is reshaping the fundamentals of production across sectors, shifting the relative importance of land toward capital and data-intensive assets[viii]. In agriculture, AI-enabled precision farming tools, including automated irrigation systems, soil and crop sensors, and drone monitoring, are increasing yields per hectare and optimising input use[ix], effectively reducing the marginal productivity of additional land in favour of technology-driven efficiency[x]. The UN Food and Agriculture Organization has documented a rising adoption of such technologies as pivotal for closing yield gaps globally[xi]. In manufacturing, smart factories powered by machine learning and robotics are enhancing design, production scheduling, and quality control, thereby diminishing traditional reliance on large industrial footprints in favour of flexible, automated facilities that leverage digital capital.

In finance and services, AI-driven platforms are expanding access through mobile and digital channels, reducing dependence on physical branches and infrastructure a dynamic accentuated in McKinsey Global Institute analyses of AI’s acceleration of financial inclusion and operational scalability. Parallel trends in urbanisation point to smart city ecosystems and digital real estate, including virtual land in metaverse environments, where economic value attaches more to connectivity and algorithms than to physical space. Collectively, these shifts illustrate the rise of technology and intellectual capital as dominant factors of production, challenging conventional assumptions about land’s pre-eminence in economic transformation, as noted in World Bank and IMF studies on AI-driven structural change[xii].

In neoclassical production terms, this reflects a shift in the Cobb–Douglas production function, where technology augments capital and raises total factor productivity, increasing substitutability between land and digital capital. As capital deepening accelerates, the elasticity of substitution between land and technology rises, weakening land’s relative factor share in value creation.

Empirical Evidence: Technology, Productivity, and the Changing Marginal Value of Land

While there is no robust empirical evidence demonstrating that artificial intelligence has directly collapsed land values as an asset class, there is substantial evidence that technology is exerting measurable “land-saving” effects by increasing output per hectare and reducing the marginal economic value of expanding land area. In China, agricultural digitisation reached approximately 27.6% of production processes by 2023, with the digital economy contributing roughly 10.5% to total agricultural value-added, reflecting a structural shift toward data-intensive farming models. Unlike South Africa, where land reform is primarily redistributive, China’s digitised agriculture model demonstrates productivity-first rural transformation, secured in technological intensification rather than ownership restructuring.

Empirical modelling in the American Journal of Agricultural Economics further estimates that, absent observed technological progress between 1991 and 2010, meeting global food demand would have required an additional 173 million hectares of cropland, an area approaching 10% of the world’s tropical rainforest[xiii]. This counterfactual illustrates how productivity gains have materially reduced the need for land expansion. At the micro-production level, precision agriculture technologies are associated with 10–30% yield improvements and significant reductions in water and fertilizer inputs, while controlled-environment and vertical farming systems can generate 10–20 times more output per unit area, using up to 95–98% less land compared to conventional cultivation[xiv]. Collectively, these findings do not suggest that land is becoming economically redundant, rather, they indicate that the marginal productivity of additional land is increasingly mediated by technology, and that future production growth may depend more on digital capital and data systems than on territorial expansion. From an endogenous growth perspective, knowledge accumulation and innovation become primary engines of sustained growth, with human capital and technological capability generating increasing returns that are not spatially bound to land. This shifts the locus of long-run development from territorial expansion to capability accumulation. For South Africa, this distinction is critical. The structural importance of land in redress and asset redistribution remains intact, but its productive dominance is increasingly conditional on technological integration.

Implications for Land Reform and Redistribution

In assessing the future of land reform in South Africa, it remains clear that land continues to be a fundamental factor of production and a structural determinant of economic inequality, even amid technological change. Decades of policy have resulted in the redistribution of only about 7–14% of agricultural land against a 30% target[xv], and land reform has had limited impact on poverty reduction and economic upliftment, in part due to under-resourcing and implementation gaps. At the same time, the rise of AI and digital technologies introduces new dynamics that both challenge and reinforce the traditional land agenda. AI-driven tools and platforms can enhance productivity and access to markets, but they also risk reducing dependence on land in some sectors (e.g., digitised services and finance) while amplifying technological divides in rural areas where broadband and digital literacy are uneven[xvi]. For example, roughly 13.6 million South Africans remain offline, largely in remote communities, despite a national internet penetration nearing 79%[xvii]. This skewed digital landscape suggests that without targeted interventions, AI could inadvertently widen the urban–rural economic divide and increase barriers for small farmers and new landowners who lack access to digital infrastructure, skills, and capital. Research from the Human Sciences Research Council cautions that AI risks deepening South Africa’s existing inequalities unless governance and innovation strategies are structured around inclusion and community capacities. Consequently, while land reform’s relevance remains strong in agriculture, mining, and real estate as bases of livelihood and wealth transfer, there is a growing imperative to complement it with AI-based rural development strategies and policy frameworks that broaden access to digital infrastructure and technological capital, thereby ensuring that technological progress does not outpace inclusion.

Looking ahead

Land reform in an AI-augmented economic environment will increasingly require hybrid approaches that integrate digital technologies with traditional land-based development pathways. AI can complement land use through precision agriculture, smart land-use planning, and data-driven resource allocation that enhances productivity per hectare, accelerating rural productivity without negating the value of land itself.

Policy recommendations emerging from this landscape include rethinking land reform frameworks to explicitly incorporate AI-driven economic models, not as replacements for land but as engines of productivity and inclusion. Fostering government private sector collaboration to bridge the AI and digital divide, ensuring that rural beneficiaries gain equitable access to both land and the technological capital required for sustainable participation in the 21st-century economy.

In conclusion, AI does not diminish the structural importance of land in South Africa’s political economy, but it does weaken its dominance as the primary driver of productive transformation, necessitating a reconfiguration of land reform policy toward a hybrid asset strategy. The central question is not whether AI replaces land, but whether South Africa can prevent technological capital from becoming the new axis of exclusion. In the 20th century, exclusion was territorial. In the 21st century, it risks becoming digital.

List of sources

[i] South African Government (2026) Land reform: historical context of the 1913 Natives Land Act and 1936 Native Trust and Land Act. Available at: https://www.gov.za/issues/land-reform (Accessed December 2025). 

[ii] South African History Online (2026) The Land Act, Traditional Authorities and the Native Affairs Department 1913–1953. Available at: https://sahistory.org.za/article/land-act-traditional-authorities-and-native-affairs-department-nad-1913-1953 (Accessed February 2026).

[iii] Grand View Research (2025) South Africa Enterprise Artificial Intelligence Market Size & Outlook, 2030. Available at: https://www.grandviewresearch.com/horizon/outlook/enterprise-artificial-intelligence-market/south-africa  (Accessed February 2026).

[iv] McKinsey & Company (2025) ‘Leading, not lagging: Africa’s generative AI opportunity’. Available at: https://aireports.africa/2025/06/12/report-on-ai-investment-trends-in-africa-2024 (Accessed February 2026).

[v] Statistics South Africa (2025) P0441: Gross Domestic Product, Second Quarter 2025. Pretoria: Stats SA. Available at: https://www.statssa.gov.za/publications/P0441/P04412ndQuarter2025.pdf (Accessed February 2026)

[vi] Minerals Council South Africa (2024) Facts and Figures 2024. Johannesburg: Minerals Council SA. Available at: https://www.mineralscouncil.org.za (Accessed February 2026).

[vii] Statistics South Africa (2025) Gross Domestic Product, Third Quarter 2025. Pretoria: Stats SA. Available at: https://www.statssa.gov.za (Accessed February 2026).

[viii] Global Growth Insights (2026) AI in Agriculture Market Size. Available at: https://www.globalgrowthinsights.com/market-reports/ai-in-agriculture-market-118527  (Accessed February 2026).

[ix] Gikunda, K. (2024) ‘Harnessing Artificial Intelligence for Sustainable Agricultural Development in Africa: Opportunities, Challenges, and Impact’, arXiv. Available at: https://arxiv.org/abs/2401.0617  (Accessed February 2026).

[x] Ken Research (2025) South Africa Digital Agriculture Platforms Market. Available at: https://www.kenresearch.com/south-africa-digital-agriculture-platforms-market (Accessed Feb 2026).

[xi] FAO (2022) The State of Food and Agriculture 2022: Leveraging Automation in Agriculture for Transforming Agrifood Systems. Rome: FAO.

[xii] World Bank (2018) Overcoming Poverty and Inequality in South Africa: An Assessment of Drivers, Constraints and Opportunities. Washington DC: World Bank.

[xiii] Villoria, N.B. (2019) ‘Technology Spillovers and Land Use Change: Empirical Evidence from Global Agriculture’, American Journal of Agricultural Economics, 101(3), pp. 870–893. doi:10.1093/ajae/aay088

[xiv] Chen, X. (2025) ‘The role of modern agricultural technologies in improving agricultural productivity and land use efficiency’, Frontiers in Plant Science, 16, 1675657. doi:10.3389/fpls.2025.1675657

[xv] MDPI (2024) ‘How Has South Africa’s Land Reform Policy Performed from 1994 to 2024?’, Land, 14(12):2443. Available at: https://www.mdpi.com/2073-445X/14/12/2443 (Accessed February 2026).

[xvi] World Bank (2018) Overcoming Poverty and Inequality in South Africa: An Assessment of Drivers, Constraints and Opportunities. Washington DC: World Bank.

[xvii] South African Government (2024) South Africa Connect. Available at: https://www.gov.za/blog/south-africa-connect (Accessed February 2026).