Load shedding has been a fact of South African business life for over 15 years, and the 2026 schedule continues to affect every business that depends on a website. A South African business that has not planned for load shedding is planning for downtime — every stage of the website's infrastructure is at risk during a power outage: the on-premise server, the ISP, the hosting provider, the office network, the content editor's laptop.
This post explains what breaks during load shedding, what business continuity looks like for a website, the planning work that should be in place, and the tooling that reduces the risk.
What breaks during load shedding
Five layers of the website infrastructure are at risk during a load shedding event:
- On-premise servers. A business with an in-office server (file server, application server, email server) loses power immediately when load shedding starts. The on-premise server goes down; any service depending on it is unavailable.
- Local network. Routers, switches, WiFi access points, VoIP phones, and other network equipment lose power. The office network goes down; staff cannot access cloud services from office devices.
- ISP infrastructure. Smaller ISPs (particularly in residential areas) have limited backup power. Even with cloud-hosted infrastructure, the office's internet connection may go down.
- Hosting provider. Major hosting providers (AWS, Google Cloud, Azure, Hetzner, Afrihost) have backup power and generator systems. The risk is low; the impact when it happens is high.
- Content editor's device. The marketing manager's laptop, the writer's desktop, the editor's tablet — all need power to function. A long load shedding event prevents the team from updating the website.
Each layer has different mitigation. A comprehensive business continuity plan addresses all five.
The business continuity impact
The business continuity impact of a website going down:
- Lead capture loss. Forms, live chat, click-to-call — all stop working. Leads are lost.
- Customer service degradation. Existing customers cannot access the website for support, account information, downloads. Service quality drops.
- SEO impact. Extended outages (4+ hours) are detectable by the engine. Repeat outages are a long-term ranking risk.
- Brand perception. Customers who cannot reach the site during a critical moment form a negative impression. The impression is hard to reverse.
- Revenue loss. eCommerce sites lose direct sales. Lead-gen sites lose indirect sales. The cost is real and measurable.
A South African business that has not planned for load shedding is exposing the business to all five impacts. The planning is not optional.
The hosting layer
The first question for any South African business: where is the website hosted, and what backup power does the hosting provider have?
Cloud hosting (AWS, Google Cloud, Azure, Hetzner, DigitalOcean, Linode):
- Major cloud providers have multi-region backup power, generator redundancy, and SLA-backed uptime commitments. Risk is low; impact when it happens is provider-specific.
- Smaller cloud providers may have less redundancy. Verify the provider's SLA and backup power claims.
South African hosting (Afrihost, Web Africa, Domains.co.za, RSAWeb):
- Most major SA hosts have backup power and generator systems. Verify the provider's track record during recent load shedding events.
- Smaller SA hosts may have less redundancy. Verify the provider's status before committing.
Self-hosted (on-premise or colocation):
- Self-hosted infrastructure is at highest risk during load shedding. The business must provide the backup power, the cooling, and the network redundancy.
- For most SMEs, self-hosting is unjustified in the South African load shedding context.
The recommendation for most South African SMEs: cloud hosting with a provider that has a strong SLA, a verified backup power system, and a track record of staying online during load shedding.
The office network layer
The office network is the second-most-common point of failure. A business with no backup power for the network equipment loses internet access within seconds of load shedding starting. Even if the website is hosted in the cloud, the office staff cannot access it.
Mitigation:
- Uninterruptible Power Supply (UPS) for the network. A UPS for the router, the switch, and the WiFi access point provides 30 to 120 minutes of backup power. The UPS is the minimum investment.
- Generator or inverter for extended outages. A business in a high-stage area should consider a generator or inverter system for the office. The system keeps the network, the desktops and the laptops charged.
- Mobile hotspot as fallback. A business with a UPS-backed router can fall back to mobile data when the fibre goes down. The fallback is slower but functional.
- Remote work as default. A business that has moved to remote or hybrid work has eliminated the office network risk. The team works from home, on their own internet connections, on devices with their own battery backup.
The office network mitigation depends on the business's reliance on the office. A remote-first business has less to mitigate.
The content editor's device layer
The marketing manager's laptop is the third-most-common point of failure. A laptop with 4 hours of battery life can survive most load shedding events; a desktop with no battery backup cannot.
Mitigation:
- Laptops over desktops. A team using laptops can work during outages; a team using desktops cannot.
- Hot desking with charging stations. A business that provides charging stations and quiet rooms for staff during outages keeps the team productive.
- Mobile devices as fallback. A team with tablets or large-screen phones can continue basic work (content editing, email, Slack) during outages.
- Hotspots for connectivity. Staff without home internet can tether to mobile networks. The connectivity is slower but functional.
The device layer mitigation is largely a function of the team's setup. A team with laptops, mobile devices and hotspot capability is resilient. A team with desktops and no mobility is fragile.
The application layer
Some websites have application-layer dependencies that fail during outages:
- Email integrations. Forms that send email notifications need the email service to be online. Major email providers (Google Workspace, Microsoft 365) have high uptime; smaller providers may not.
- CRM integrations. Forms that push to a CRM need the CRM to be online. Most major CRMs have high uptime.
- Payment gateways. eCommerce sites need the payment gateway to be online. Major gateways (PayFast, PayGate, Stripe, Yoco) have high uptime.
- Live chat. Live chat services need to be online. Most major services have high uptime.
- Third-party APIs. Any third-party API call (currency conversion, shipping calculator, address validation) needs the API to be online.
The application layer risk is provider-specific. The mitigation is to choose providers with strong uptime and to design the application to degrade gracefully when a dependency is offline.
The planning work
For a South African SME, the load shedding planning work:
- Audit the infrastructure. Document the hosting provider, the office network, the team's devices, the application dependencies. Identify the highest-risk layer.
- Verify the hosting provider. Confirm the provider's backup power, the SLA, the track record. Switch providers if necessary.
- Invest in a UPS. A UPS for the office network is the minimum. A larger UPS or generator system for extended outages is worth the investment for high-stage areas.
- Equip the team. Laptops, mobile devices, hotspot capability. The team should be able to work from anywhere during an outage.
- Document the failover plan. When load shedding starts, what does each person do? The plan should be documented, rehearsed, and accessible offline.
- Test the plan. Quarterly load shedding fire drill — simulate an outage and verify the team can continue working. The test catches the gaps.
A business that completes the planning work is resilient to load shedding. A business that has not done the planning work is exposed to outages 4+ hours at a time, several times per week in high-stage periods.
The cost of the planning work
The planning work has real costs:
- UPS. R1,500 to R10,000 depending on capacity.
- Generator or inverter. R15,000 to R100,000+ depending on capacity.
- Laptops over desktops. R8,000 to R25,000 per device.
- Mobile hotspots and data. R500 to R2,000/month per user.
- Hosting upgrade. R500 to R5,000/month extra for higher-tier cloud hosting.
The total cost for a small team (5 people): R50,000 to R150,000 for hardware, R5,000 to R15,000/month for ongoing. The cost is justified when the website is core to revenue.
The alternative: remote-first
A South African business that adopts remote-first work eliminates most of the load shedding risk:
- No office network to fail. Each team member uses their own home network and power.
- Each team member is responsible for their own backup. Personal UPS, personal hotspot, personal laptop battery.
- The hosting provider is the only single point of failure. With a strong cloud provider, the risk is low.
The remote-first model is the most resilient to load shedding. The cost is the cultural shift and the management overhead of distributed work. The benefit is the load shedding resilience.
FAQ
How long can a UPS power a network?
A 1,500 VA UPS powers a typical home/office network (router + switch + WiFi + small NAS) for 30 to 90 minutes. A larger UPS powers the network for 2 to 4 hours.
Should I host my website with a South African provider?
Not necessarily. Major international cloud providers (AWS, Google Cloud, Azure, Hetzner) have stronger uptime guarantees than most SA providers. The "SA provider for SA audience" argument is weak because the audience does not care where the site is hosted, only that it loads fast.
Can a generator keep a website online?
Yes, but only if the generator powers the office network, the team's devices, and (if self-hosted) the server. A generator for the office network alone does not keep the website online if the hosting provider is the issue.
How do I know if my hosting provider has backup power?
Ask them. Look for the SLA in the contract. Check reviews from other customers during load shedding events. The information should be available; if it isn't, the provider is not transparent.
What about a CDN — does that help with load shedding?
Yes, indirectly. A CDN caches the site at edge locations around the world. If the origin server goes down briefly, the CDN can serve cached pages. The protection is partial; dynamic content still requires the origin. But the CDN adds resilience.
If you want help auditing the load shedding readiness of a specific business, send us the hosting setup, the office network, the team's devices, and the application dependencies. Most business continuity audits we scope are 2 to 4 hours of work, with the highest-risk layer identified and the mitigation scoped separately.