South Africa has been manufacturing products from food to locomotives (several manufacturers but killed off because of policy) and, as is well known, motor vehicles. Its range of manufacturing is tremendous, and partly what explains that is it’s better to manufacture locally than import, because of the huge distances that South Africa has to major European and other markets.

Some manufactured goods in South Africa are world-class, and I read a report recently where a kiln manufacturer in Alrode, near Johannesburg, had most of its products exported to Australia. Then a few years ago I was at a function where a manufacturer from Potchefstroom explained how 90% of his production was exported.

Let’s look at the manufacturing index in South Africa, then look at some of the challenges, including the well-known ones, but ones like, for example, inside the factory, which includes the rising costs within the factory, the undercapacity and underutilization, the need to renew manufacturing machinery and infrastructure in the plant. We’ll look at the Internet of Things and touch on not just preventive maintenance, but rather predictive maintenance and how AI can help with predictive maintenance.

South Africa’s manufacturing sector is still under pressure, but the latest data shows a modest improvement, with the manufacturing production index up 1.1% year on year in July 2026 after earlier declines.

Sectors doing well

The strongest support came from food and beverages, which rose 4.1%, and from petroleum, chemical products, rubber and plastic products, which increased 3.2% and made one of the biggest positive contributions to overall manufacturing output. Stats SA also noted that the petroleum, chemical products, rubber and plastic products division was a positive contributor in the second quarter, alongside motor vehicles, parts and accessories and other transport equipment.

The broader manufacturing PMI still suggests weakness in factory conditions, with the index at 45.8 in August, below the neutral 50 mark.

Manufacturing’s recent performance in South Africa has been a study in contrasts: pockets of excellence and export success sit alongside a broadly sluggish index. Output growth, where it exists, is modest and uneven across sectors. Capacity utilisation remains stubbornly below historical norms, and the capital stock is ageing. Those macro numbers mask a layered problem that runs from ports and power grids down to the grease on the factory floor.

The usual culprits — energy insecurity, logistics bottlenecks, and policy uncertainty — deserve every airing. They raise costs, shave margins and make long-term investment decisions fraught. But equally important and often overlooked are the pressures originating inside plants. Rising input and labour costs are amplified by inefficiencies rooted in ageing machines, patchy maintenance regimes and production lines designed for a different competitive era. Underutilised equipment and ad hoc repairs are expensive: they nudge firms toward extra inventory, firefighting, and sometimes lost export contracts when delivery windows are missed.

Renewal of plant and machinery is costly and discretionary. Many firms postpone capex because of uncertain demand and tight balance sheets. The result is a vicious cycle — older machines break more often, maintenance becomes reactive and downtime rises, which suppresses output and cash flow, further delaying investment.

Here a pragmatic lever exists: smarter maintenance powered by the Internet of Things (IoT) and artificial intelligence. Moving beyond calendar-based preventive maintenance to condition-based and predictive maintenance can materially reduce unplanned downtime. Sensors fitted to motors, bearings, temperature controls and conveyors stream continuous data. Machine-learning models ingest that data to detect subtle patterns — rising vibration, microtemperature shifts, or anomalous current draws — that foreshadow failure. Alerts trigger targeted interventions, not blanket shutdowns.

The benefits are multiple and immediate. Predictive maintenance extends asset life, reduces spare-part inventories and lowers labour cost per output unit because technicians are deployed efficiently. It also improves reliability, a simple but powerful sales argument for local manufacturers competing in export markets that prize on-time delivery. For smaller plants, cloud-based IoT solutions and pay-as-you-go analytics lower the entry bar; they don’t need to own data centres or hire teams of data scientists to get started.

Adoption is not automatic. Firms must overcome skills gaps, integrate new data streams into legacy control systems, and make modest upfront investments in sensors and connectivity. Regulators and industry bodies can help by promoting standards for interoperability and targeted incentives for digital retrofits. Banks and leasing houses could tailor finance products for incremental capex on digital upgrades rather than large-scale greenfields.

South African manufacturing faces a multi-front challenge, but the remedy need not be grand or expensive. Anchoring competitiveness in better machine health and smarter operations is a practical, high-return move. It won’t erase structural problems like power shortages or global demand shocks, but it can strengthen margins, preserve export credibility and buy time for the bigger reforms and investments that manufacturing ultimately needs.

Editorial Disclosure & Disclaimer

Financial News Daily is an independent business news syndicate and a wholly owned subsidiary of Idea Accelerator. We publish financial, environmental and corporate commentary for digital platforms, media outlets and organisations. Nothing published here constitutes investment, legal or financial advice. All opinions are editorial commentary on matters of public and economic interest.