How AI is rewiring South Africa

Despite political uncertainty, escalating crime and a struggling economy, South Africa has emerged as one of Africa's most advanced AI ecosystems. The country combines the continent's deepest enterprise market, its most developed cloud infrastructure, top universities and the presence of the world's major global technology companies, among them Google, Microsoft and Amazon. Over the past 18 months the South African AI conversation has shifted from experimentation to large-scale deployment across banking, retail, healthcare, mining and government. Other factors underpinning the country's AI ecosystem include a mature financial services sector willing to invest in AI, strong software engineering talent and major cloud infrastructure from US tech giants. South Africa also boasts world-class research institutions such as the Council for Scientific and Industrial Research (CSIR), the University of Cape Town, the University of the Witwatersrand and Stellenbosch University. From pilot projects to full-scale deployment Industry analysts increasingly describe 2026 as the year South African companies moved from AI pilots to operational deployment, particularly in customer service, coding, cybersecurity, fraud detection and productivity. Major South African AI players such as Entelect, BBD, Dariel, EOH (originally Enterprise Outsourcing Holdings) and BCX (Business Connexion) are integrating generative AI into software development, banking, telecommunications and public-sector projects. In addition, South African banks are among Africa's most advanced AI adopters. These include FirstRand, Standard Bank, Nedbank and Absa. The country's highly competitive retail sector has also become one of the fastest-growing AI players in the country. Supermarket giant Pick n Pay recently launched its "Penny" campaign, an AI shopping assistant built on Google's Gemini models. Rival Shoprite has expanded its AI-powered personalised shopping and replenishment tools. Attractive data centre picture The country now hosts cloud regions from Google Cloud, Microsoft and Amazon Web Services. This dramatically reduces latency - the time it takes for a user to receive a response to an online request. At the same time it helps organisations to comply with data-residency requirements. When it comes to data centres, the country is host to major operators, including Teraco, Africa Data Centres, Open Access Data Centres and NTT DATA. Johannesburg remains Africa's largest data-centre hub, despite the current parlous state of its infrastructure. South Africa also benefits from multiple international submarine cables, including 2Africa submarine cable, Equiano, WACS and SEACOM. James Manyika, a senior partner emeritus at McKinsey & Company and chairman and director emeritus of McKinsey Global Institute (MGI) says the world is seeing AI adoption accelerate at an extraordinary pace. Africa, he says, "now has an opportunity to move beyond merely using AI and towards building AI companies". Manyika says AI is ushering in "a new global engine of innovation powered by African talent". He believes South Africa is a crucial launchpad for the continent, and stresses that AI acts as an economic catalyst that will transform industries, improve public services and create new jobs. A general view of the MeerKAT radio control room in the South African Radio Astronomy Observatory (SARAO) headquarters in Cape Town. (Photo by MARCO LONGARI / AFP) Regulatory environment vacuum Unlike the European Union (EU), South Africa has not yet formulated any meaningful dedicated AI legislation. Instead, AI is currently governed through existing, pre-AI, laws. These include the Protection of Personal Information Act (POPIA), consumer protection legislation, copyright law, cybercrime legislation and sector-specific financial regulations. The government intended to release a Draft National Artificial Intelligence Policy Framework that sought, among other things, to establish ethical AI principles, promote responsible deployment, develop AI skills and strengthen AI governance. This framework sought to avoid imposing the highly prescriptive obligations found in the EU AI Act, instead favouring a phased, risk-based approach. However, the draft policy was withdrawn after AI-generated citations were found. The target publication date for public comment is now January 2027. President Cyril Ramaphosa recently argued that South Africa must balance innovation with reliable safeguards. "There is a need for guardrails to prevent abuse and other risks," he said. The president stressed that data sovereignty would become a strategic national issue as AI infrastructure expands. "Our country's ambition to be a leading digital investment destination is being recognised by some of the world's leading technology companies," he said. Ramaphosa also warned that "innovation must be matched by appropriate regulation." Google ramps up presence In the meantime Google has become one of the country's largest AI investors. At its inaugural Africa Cloud Summit in Johannesburg in July 2026, Google announced that it had exceeded its $1bn African investment commitment and unveiled several South African initiatives. These include the expansion of AI infrastructure; an applied AI lab, the expansion of the Google for Startups accelerator programme for South Africa; the introduction of AI skills programmes; free Gemini AI access for university students; and additional cloud investment around the Johannesburg Cloud Region. Google estimates that its Johannesburg Cloud Region could contribute roughly 1.7 trillion rand ($1bn) in additional economic output and claims that it will support around 315,000 jobs by 2030. The corporation also recently announced an additional digital exchange port in the Eastern Cape, linked to its global subsea network, strengthening South Africa's role as an AI and cloud gateway between Africa and Asia-Pacific. The Google Cloud logo is pictured at the MWC (Mobile World Congress), the world's biggest mobile fair, in Barcelona. (Photo by Manaure Quintero / AFP) Businesses look to opportunities Enrico Ferigolli, retail executive - omnichannel at retail giant Pick n Pay, leads the company's rollout of AI-powered features. He says that AI in South Africa is changing how consumers shop. "While on-demand delivery transformed how South Africans buy groceries, AI is now changing how they order," he says, describing conversational AI as "the next major shift in online retail". He explains that the country's first conversational AI grocery shopping assistant, Penny, integrated directly into the Pick n Pay app and powered by Google's Gemini AI models, allows shoppers to build entire baskets, plan meals and request product swaps using voice notes, text, or photos. Ferigolli notes that Pick n Pay evaluated several large language models before selecting Google Gemini, due to its balance of accuracy, speed and cost. He says multilingual capabilities - particularly for South Africa's African languages - are still improving, although voice support is more advanced than text. Consumers, Ferigolli says, "no longer just want speed; they want shopping apps to think for them". Doing more with less Benjamin Rosman, professor at the School of Computer Science and Applied Mathematics at the University of the Witwatersrand, tells African Business that AI is likely to be deeply consequential for South Africa, but its impact is not predetermined: "AI could improve business productivity, public services, infrastructure planning and access to expertise, but only if we build the capacity to shape and govern it locally. Otherwise, there is also the potential for it to widen inequality and become a destabilising force in society". AI can help South Africa "do more with limited resources," Rosman says. "The most important opportunities are in areas such as healthcare, education, agriculture, public administration, scientific discovery, financial inclusion and service delivery. For ordinary South Africans, the promise is better access to expertise, faster services, more personalised support and new forms of economic participation. "South Africa needs to treat AI as a national capability, not just a set of imported tools. This means investing in education, research, compute access, data governance, public-sector capacity and local innovation. "Most importantly, we need to ensure that AI is designed and deployed around South African realities, rather than assuming that systems built for other contexts will automatically serve us as well. This requires tackling the topic as a key priority in the country," he observes. Civil society, Rosman adds, has a crucial role to play in ensuring that AI serves people, protects rights and does not deepen existing inequalities. He says the biggest risks are job displacement, concentration of power, misinformation, cybersecurity threats and increased dependency on foreign technologies. "In a country with high unemployment and deep inequality, these risks matter enormously. We should not respond by rejecting AI, but by steering it deliberately and building the institutions needed to govern it". A general view of the Alexandra township with the Sandton cityscape in the background taken from Linbro Park in Johannesburg. (Photo by EMMANUEL CROSET / AFP) The challenges that lie ahead Despite its African leadership position in AI, South Africa faces significant constraints. Electricity shortages remain a concern for data centre expansion, although reliability has improved significantly since 2024, when the country was beset daily by what was euphemistically known as "load-shedding", the bane of every South African's life. Adding to this, AI adoption is concentrated among large enterprises, with slower uptake by small and medium enterprises (SMEs). Policymakers continue to debate issues of data sovereignty, intellectual property, bias and the impact of AI on employment. Nevertheless, South Africa is increasingly positioning itself as Africa's principal AI hub. Its advantages lie less in creating frontier AI models than in providing the continent's most mature environment for deploying AI at scale. South African businesses, government and civil society increasingly see artificial intelligence not simply as a productivity tool, but as an opportunity to address long-standing structural challenges - from low productivity and skills shortages to healthcare access, multilingual communication and financial inclusion. Industry bodies and technology companies generally identify several areas where AI could have the greatest impact in South Africa. These include productivity gains across finance, mining, retail, manufacturing and professional services, public service delivery including healthcare diagnostics, education and citizen services and small business growth through affordable AI-powered marketing, accounting and customer service. Crucially, they also incorporate making digital services accessible to millions previously excluded by English-only technology. The question, it seems, is no longer whether artificial intelligence will reshape South Africa, but whether South Africa can move quickly enough to reshape AI for the benefit of all its people - and persuade it to talk the right language.