South Africa’s artificial intelligence market, already worth more than R50 billion, is on track to triple in size by 2030, according to figures presented by Huawei corporate vice president Nicholas Ma and reported by ITWeb. The projection puts the country at an inflection point where the difference between capturing that growth locally and watching it flow to better-prepared markets elsewhere will come down to infrastructure decisions being made, or delayed, right now.
Ma’s presentation, delivered alongside Huawei South Africa chief executive Will Meng and attended by Communications and Digital Technologies Minister Solly Malatsi, cited research suggesting AI could contribute between R1 trillion and R1.4 trillion to South Africa’s GDP by 2030, alongside the creation of 200,000 to 400,000 ICT jobs over the same period. Those are the kind of headline figures that get cited in policy documents long after the underlying assumptions have been forgotten, but the more grounded numbers in Huawei’s presentation are arguably more useful: 90% of surveyed working professionals already use AI tools during a typical work week, a figure that suggests adoption has moved well past the early-adopter phase and into something closer to baseline expectation.
The infrastructure bottleneck nobody wants to own
Globally, Huawei’s figures point to roughly 30 million active AI agents already deployed, with a prediction of 90 billion agents by 2035 at an assumed 85% adoption rate, a scale of deployment that Gartner separately estimates could generate $4 trillion in economic value worldwide over the next five years. Whatever share of that South Africa actually captures depends less on enthusiasm, which the 90% weekly-usage figure suggests is not in short supply, and more on unglamorous infrastructure: resilient networks, cloud platforms and data centre capacity able to support autonomous AI systems operating at scale, rather than the chatbot-style tools most businesses have adopted so far.
That is a harder sell politically than a jobs number, since data centre capacity and grid-reliable power for compute-hungry AI workloads compete directly with a national electricity system that has spent the better part of a decade being rebuilt from a much lower base. South Africa’s own draft national AI policy has been through a long consultation process without yet landing on binding rules, and the gap between having an AI strategy on paper and having the electricity, fibre and data centre capacity to run it at scale is precisely the gap Huawei’s presentation was aimed at.
The recommendations that came out of the session read like a checklist rather than a strategy: build computing power for autonomous systems, integrate AI into everyday business workflows, break down the data silos that keep South African enterprises from pooling the information their own AI systems would need to be useful, close the AI skills gap, and expand affordable device access and connectivity so adoption is not confined to companies that can already afford enterprise-grade infrastructure. None of it is controversial. All of it requires sustained investment at a moment when government’s own fiscal space is already stretched thin by more immediate priorities.
Huawei’s own commercial interest in the outcome is worth naming plainly. The company sells the networking equipment, cloud infrastructure and data centre hardware that a tripling AI market would need far more of, which means Ma’s presentation functions simultaneously as a market forecast and a sales pitch for exactly the kind of infrastructure Huawei builds. That does not make the underlying adoption figures wrong, but it is a reason to treat the most bullish GDP contribution numbers, the R1 trillion to R1.4 trillion range in particular, with the same scepticism due any projection produced by a company that stands to benefit directly from policymakers acting on it.
What is harder to dismiss is the skills and jobs dimension, since a market tripling in size without a matching increase in AI-literate workers simply means South African companies end up paying foreign contractors and platforms to capture value that could otherwise stay local. The 200,000 to 400,000 ICT jobs figure cited alongside the GDP projection depends entirely on training pipelines, from universities through to vocational programmes, actually producing workers with the specific skills an AI-driven economy needs, rather than a general computer science graduate who has never touched a production machine learning system. That training gap, more than compute capacity or even electricity supply, is the constraint least likely to be solved quickly, since it takes years to build a cohort of graduates rather than months to commission a data centre.
What makes the tripling projection worth taking seriously rather than filing alongside every other bullish five-year forecast is the adoption figure sitting underneath it. A market does not typically triple on the back of infrastructure investment alone, it triples when businesses have already decided the technology is worth paying for and are simply waiting for the capacity to use more of it. If nine in ten working South Africans are genuinely already using AI tools weekly, the demand side of Huawei’s forecast looks less like marketing and more like a market that is already ahead of the infrastructure meant to serve it, which is an unusual, and arguably more investable, place for a national tech sector to find itself. That gap between enthusiasm and the rules meant to govern it is also visible closer to home: see this site’s report on a recent cybersecurity breach that exposed gaps in South Africa’s own AI policy framework.
