# Hosting and Load Shedding: Uptime Strategy for South African Businesses
Load shedding is currently suspended, with Eskom's grid stable for over 300 consecutive days as of early 2026 — but the underlying capacity gap hasn't gone away, and analysts still flag real risk of recurrence through 2026–2030. A website hosted properly doesn't actually depend on the answer either way, which is the point of this guide.
The Honest Starting Point
It's tempting to write this as a fear-based pitch — "load shedding could kill your website, act now" — but that overstates the current risk and undersells the real one. Right now, load shedding isn't happening. What's still true, load shedding or not, is that South African businesses lose more uptime to ordinary causes — fibre cuts, ISP outages, a server misconfiguration, a DDoS attempt — than to Eskom specifically. A resilience strategy built only around Stage 6 blackouts misses the more common failure modes. Build for both.
How Eskom's Stages Actually Work
Load shedding is measured in stages, and the stage number tells you how much national capacity needs to be shed, not directly how many hours you'll lose — though higher stages mean more frequent and longer cuts in practice.
| Stage | Capacity Shed | Typical Impact |
|---|---|---|
| Stage 1 | Up to 1,000MW | A few 2–4 hour cuts over several days |
| Stage 2 | Up to 2,000MW | More frequent cuts, same duration pattern |
| Stage 4 | Up to 4,000MW | Cuts most days, often twice daily |
| Stage 6 | Up to 6,000MW | Up to around 12 hours of cuts a day, in 2–4 hour blocks |
| Stage 8 | Up to 8,000MW | The most severe published stage — extended, frequent outages |
The practical business question isn't "what stage are we at" — it's "does our hosting depend on power staying on at one specific building." For most small SA businesses using shared or basic hosting, the honest answer used to be yes.
Where Your Website Actually Lives, and Why It Matters
A website's uptime during a power event depends on three separate things, and only one of them is in your control:
1. The hosting server's power supply — is the data centre it sits in on generator/UPS backup, or does it go down when the building's power does?
2. Your visitors' own connectivity — you can't fix a customer's local outage, but you can make sure your site loads fast and light so it works on their phone's data even when things are strained
3. Your own ability to operate — orders coming in, support queries, anything requiring you personally to respond
Most South African hosting conversations only address the first, but the second matters just as much for e-commerce specifically: a slow-loading site or a payment gateway timeout during a period when everyone's connection is already stressed is a direct, measurable loss of sales — arguably a bigger revenue risk than the server itself going down.
What Real Infrastructure Redundancy Looks Like
South Africa has genuinely world-class data centre infrastructure at the top end, which is worth knowing about even if you're on a modest hosting plan, because it explains what you're paying for when a host charges more:
- [Teraco](https://www.teraco.co.za) operates Africa's largest carrier-neutral data centre campus, at Isando in Johannesburg, and provides direct cloud on-ramps for AWS, Google Cloud, Microsoft Azure, and Oracle Cloud. Its facilities are built to a 99.999% uptime design standard.
- Xneelo, one of SA's longest-running hosting providers, runs on a 2N fully redundant power system with physically separate A+B power feeds and diesel generators that can run indefinitely if utility power is lost.
- Afrihost, one of the most recognisable SA hosting brands, hosts out of Teraco's Isando campus rather than running its own facility — which is a completely standard and sensible approach, not a shortcut.
The pattern across all of them: the data centre itself is essentially load-shedding-proof, because generator and UPS redundancy is the entire point of a proper Tier III facility. The risk isn't usually "will the data centre lose power" — it's "is my hosting actually in one of these, or on a cheaper setup without that backup."
A Practical Uptime Checklist
- Confirm your host's actual power redundancy, not just their uptime percentage claim — ask directly whether their facility has generator backup and what "99.9% uptime" is measured against
- Put a CDN in front of your site (Cloudflare's free tier is a reasonable starting point) so cached pages keep serving even during a brief origin hiccup, and so international or out-of-province visitors aren't routed through a single Johannesburg link
- Keep pages light — this matters every day, not just during load shedding, but it matters more when visitors are on strained mobile data during an outage
- Have a status page or a simple "we're aware, here's how to reach us" fallback for your support channels, so a genuine outage doesn't also look like radio silence
- Don't rely on a single ISP link for anything business-critical you run yourself (a local server, an on-premise POS system) — that's a separate risk from your website's hosting, but it's the one most SA businesses actually feel first
- Test your backups, not just that they exist — a backup you've never restored from is a hope, not a plan
FAQ
Is load shedding actually a risk in 2026, or is this outdated advice?
As of early 2026, load shedding has been suspended for over 300 consecutive days — but Eskom's underlying capacity constraints haven't been resolved, and independent analysts still flag risk of recurrence later in the 2026–2030 window. Building resilience now costs little and isn't wasted effort either way, since the same measures protect against ordinary outages too.
Does hosting my site outside South Africa avoid the problem entirely?
It removes the local power dependency, but it adds latency for South African visitors and doesn't help with the more common causes of downtime (ISP issues, DNS problems, application errors). A CDN in front of local hosting usually gets you most of the benefit without the latency cost.
What actually happens to my website during a power cut if my host has generator backup?
Nothing visible to a visitor — that's the entire point of the redundancy. The generator kicks in within the facility's designed failover window (typically seconds, not minutes) and the server keeps running.
Is shared hosting ever good enough, or do I need a dedicated/cloud setup?
For a low-traffic brochure site, reputable shared hosting from a provider with real backup power is fine. The moment you're taking payments, running an app customers depend on, or getting meaningful traffic, the calculus changes — not because shared hosting is inherently unreliable, but because the cost of downtime rises faster than the cost of better infrastructure does.
How do I check if my current host actually has proper power redundancy?
Ask them directly, in writing: which data centre they use, whether it has generator backup, and what their actual (not marketed) uptime has been over the past 12 months. A host that can't answer specifically is itself the answer.
---
If uptime and performance are a genuine business risk for you — not just a nice-to-have — that's core to how we architect and host projects. See our cloud and DevOps work or get in touch to talk through what your current setup actually depends on.