What is the best backup power for a small business during load shedding?

For most offices, a pure sine-wave inverter with lithium batteries sized to your router, switches, laptops and lighting is the best balance of cost and run-time. Generators suit heavy loads but cost more to run; UPS units are for bridging seconds, not hours. Size the system on your actual watt draw, not guesswork.

How do I keep internet during load shedding?

Put your router and ONT on the inverter or a dedicated mini-UPS, and add a second connection on a different medium — fibre plus a 5G/LTE failover from a different provider. Towers and exchanges can also go down, so diversity of medium matters more than speed.

Should I move my business systems to the cloud because of load shedding?

Yes for anything customer-facing or critical. Cloud-hosted systems keep running when your office is dark — your team can work from anywhere with power. On-premise servers turn every stage of load shedding into downtime and risk hardware damage from dirty power.

← SIGNAL · Resilience

Load shedding & your business: the 2026 IT continuity playbook

By KTH-Tech · Resilience & infrastructure · 9 min read

Load shedding is no longer an event in South Africa — it's an operating condition. The businesses that thrive are the ones that engineered for it: power that fails over silently, connectivity that survives a dark tower, and systems that don't live in the office at all. This is the practical playbook we use when we design continuity for clients — sized for SMEs, applicable to enterprises.

You can't control the grid. You can control whether stage 6 costs you revenue, data, or customers.

1.Power: size it on real numbers

Most businesses buy backup power backwards — a generator "that should be enough". Start by measuring what actually must stay on: router and ONT (10–25W), a switch (10–30W), laptops (30–65W each), lighting, card machines. A team of ten's critical load is often under 800W — that's inverter-and-lithium territory, not diesel.

UPS — bridges seconds to minutes. For servers and desktops so they shut down cleanly, not for running the day.
Inverter + lithium battery — silent, instant switchover, no fumes, scales from a router-only mini unit (~R2,500) to whole-office systems. The default answer for offices.
Generator — for heavy loads (kitchens, workshops, HVAC). Budget for fuel, servicing and a changeover switch; never run IT directly off one without a UPS smoothing the power.
Solar + storage — the long-term play. Load shedding pays back the battery; sunshine pays back the panels. Get a CoCT/SSEG-compliant installer and the paperwork done.

2.Connectivity: diversity beats speed

When the lights go out, fibre usually stays up — but the tower or exchange serving you may not. The pattern that works is two connections on two different mediums from two different providers: fibre as primary, 5G/LTE failover on a router that switches automatically. Keep the ONT and router on protected power (they draw almost nothing), and check your failover actually fails over — test it during the next stage, not during a client demo.

3.Systems: get them off the premises

An on-premise server during load shedding is downtime plus hardware risk from dirty power. Anything customer-facing — website, ordering, bookings, support — should live in the cloud, where a properly built site keeps trading while your office is dark. Same for the tools your team uses: cloud accounting, cloud files, cloud email mean "work from anywhere with power" is a plan, not a scramble.

4.Data: assume the worst day

Power cuts corrupt disks and interrupt syncs. Your protection is the boring, non-negotiable 3-2-1 rule: three copies of critical data, on two different mediums, one off-site (cloud). Automate it — a backup that depends on someone remembering is not a backup. And test a restore quarterly; plenty of businesses discover at the worst moment that their backups were empty.

5.People: write the two-page plan

Continuity fails at handover, not hardware. Write a two-page plan everyone knows: who switches to failover, where the team works during extended outages, how customers reach you when the office line is down (a cloud PBX or WhatsApp Business number), and who talks to clients if systems degrade. Rehearse it once — a 30-minute dry run finds the gaps a document can't.

The stage-6 survival checklist

Can your business say yes to all of these?

Critical load measured in watts, and backup power sized to it

Router + ONT on protected power with automatic switchover

Second internet connection on a different medium and provider

Customer-facing systems hosted in the cloud, not the office

Automated 3-2-1 backups, with a restore tested this quarter

A two-page continuity plan the whole team has actually seen

The bottom line

Load shedding punishes improvisation and rewards engineering. The stack above — right-sized power, diverse connectivity, cloud systems, automated backups, a rehearsed plan — is achievable for an SME in a month, and it turns stage 6 from a crisis into a non-event. If you're modernising systems anyway, start with our guide to AI readiness — resilience and intelligence are built on the same foundations.

Want continuity engineered, not improvised?

KTH-Tech designs load shedding-proof infrastructure and cloud platforms for South African businesses — sized to your real load, budget and risk.

Talk to an engineer →

General guidance, not electrical advice. Installations must be done by certified electricians and comply with your municipality's SSEG rules.