Data, Devices and Load Shedding: What's Actually Still Stopping AI Adoption in 2026

This is worth stating clearly and directly, because the assumption persists long after the reality has shifted: as of May 2026, South Africa had gone a full year without a load shedding event, according to African Business reporting. This represents a genuine and significant change from the disruption patterns of previous years.

To be fair to the historical record, load shedding was a real and well-documented barrier to technology adoption for informal businesses. A 2024 study on informal businesses in Cape Town specifically identified load shedding among the concrete barriers blocking technology adoption at the time. But treating it as a current, ongoing constraint in 2026 content or planning is now out of date — the live infrastructure conversation in South Africa has shifted toward grid theft and infrastructure vulnerability rather than scheduled outages.

The more current and relevant constraint is data affordability — and the picture here is genuinely mixed, showing real progress alongside a real ongoing gap.

ICASA's 2026 State of the ICT Sector Report shows South Africa's entry-level mobile broadband basket improved to R152 per month in 2024, equal to 1.63% of monthly GNI per capita. This technically places South Africa below the widely-used 2% global affordability benchmark — a genuinely positive development.

However, the same ICASA report notes that this figure still represented the highest entry-level mobile broadband cost among BRICS countries. So while South Africa has crossed a general affordability threshold, it remains comparatively expensive relative to peer economies — a nuance that matters for a business owner deciding whether frequent AI tool use is a viable regular expense.

On smartphone access specifically, ICASA reported the lowest entry-level smartphone price in 2025 at R399 — a meaningful drop that lowers the barrier to entry for a first smartphone or an upgrade to a device capable of running modern AI apps effectively.

Older but still relevant research from Research ICT Africa found 93% mobile phone penetration among owners of informal and small businesses specifically — showing that basic device access has been high for some time. The more telling finding from the same research: less than a fifth of these business owners were using their mobile phones for actual business processes and management, as opposed to personal use. This points to a usage and skills gap sitting alongside — and arguably larger than — the device access gap itself.

Pulling this together honestly: the infrastructure barriers most commonly assumed — no electricity, no phone, no internet at all — are considerably less accurate as a picture of 2026 reality than they were even a few years ago. Devices are cheaper. Data has crossed a global affordability benchmark, even if it remains comparatively expensive regionally. Load shedding, while not forgotten, is not currently disrupting daily operations.

What the research consistently points to instead is a different kind of gap: data cost remaining meaningful for low-margin businesses even after improvement, and — more significantly — a gap between simply having a device and internet access, and actually knowing how to use that access for structured business purposes rather than personal communication alone.

This matters practically because it changes where effort and support should be focused. If the primary barrier were purely infrastructure — devices, connectivity, electricity — the solution would be largely about cost and access. But if the primary barrier is increasingly about skills, structured use, and knowing specifically how to apply available tools to business problems, the solution looks different: training, guided application, and hands-on support matter more than simply providing cheaper access to technology that's already reasonably accessible.

This distinction directly shapes kasiAIhub's approach. Rather than assuming the barrier is primarily about access — most participants already have a smartphone and some data — the programme focuses on the actual gap the research identifies: structured, hands-on guidance in applying AI tools to specific business functions, building the skill and habit of business-purposed use on top of infrastructure that, in most cases, already exists.

See how hands-on training addresses this gap across Think & Plan, Build & Create and Brand & Grow. This piece sits inside the broader State of AI Adoption Among Township Entrepreneurs (2026) hub.

See the full 3-session programme →

African Business reporting on load shedding (May 2026); South African Journal of Information Management research on Cape Town informal business technology barriers (2024); ICASA State of the ICT Sector Report (2026); Research ICT Africa mobile phone penetration research.