Telstra Outage: Protect Your Business From the Next One

In the early hours of Wednesday 8 July 2026, Telstra’s mobile network started failing — and by the time most business owners had made their first coffee, the impact was national. Mobile calls dropped or wouldn’t connect. Data came and went. EFTPOS terminals in cafes, clinics and shops stopped taking payments. Train services were delayed in Victoria. Most seriously of all, some triple-zero calls didn’t connect, triggering hundreds of urgent welfare checks. By 7am, DownDetector had logged more than 7,000 outage reports.

Telstra traced the fault to a software defect. From around 4:30am AEST, network nodes responsible for synchronising time across its mobile network stopped operating as expected, causing intermittent failures anywhere in the country depending on how traffic was routed. It wasn’t a cyberattack. Most services were restored by 10am — just under 90 per cent of calls and data working normally — with full recovery later that day, and the ACMA has launched a full investigation.

The post-mortem is Telstra’s problem. Yours is simpler and more uncomfortable: if your business runs on one carrier, one connection and one router, you don’t have bad luck when this happens — you have a single point of failure with a logo on it. The good news is that outage resilience is a design problem, and design problems can be solved. Here’s how.

What actually happened — and why it hit businesses so hard

Mobile networks depend on extremely precise timing to coordinate calls and data sessions between towers and core systems. When the nodes managing that synchronisation misbehave, the network doesn’t fail cleanly in one region — it fails intermittently, everywhere. That’s exactly what businesses experienced on Wednesday: two phones side by side could behave completely differently, and a call that failed at 8am might work at 8:05.

The bigger lesson is the blast radius. A modern business quietly rides the mobile network in far more places than the phones in staff pockets: SIM-based EFTPOS terminals, back-to-base alarms, on-call and duress phones, courier and dispatch apps, remote monitoring sensors, and the 4G backup links buried inside other equipment. Telstra itself acknowledged that payment systems were collateral damage. Two practical postscripts are worth noting for your own playbook: some business devices needed a manual restart even after the network recovered, and Telstra warned customers to hang up on unsolicited calls about the outage, because scammers move fast on days like this.

This isn’t a one-off. It’s a pattern.

If Wednesday felt familiar, that’s because Australia has now had a major carrier failure roughly once a year for four years running — and each one caught thousands of businesses with no fallback.

The consumer watchdog ACCAN’s response to this week’s outage was blunt: Australia still has no enforceable reliability standards holding telcos to account for network stability. Translation for business owners — no regulator is about to make outages impossible. The only network you control is your own, and it can be designed to survive your carrier failing.

What a morning offline actually costs

Downtime maths is unforgiving because the meter runs on several things at once: staff you’re paying who can’t work properly, sales you can’t take, calls that ring out to a competitor, and the recovery hours afterwards. Run your own numbers with the formula below — most businesses are surprised at how quickly it climbs.

For regulated businesses, the cost isn’t only measured in dollars. An aged care provider whose on-call escalation number rings out has a clinical risk, not an IT inconvenience. A financial advice firm unreachable during market hours has a client duty problem. A law firm that misses a court registry deadline has an explanation to write. The dollar figure is just the part that’s easy to add up.

The fix: build layers, not luck

Real continuity isn’t one product you buy — it’s a stack of unglamorous decisions that each remove a single point of failure. Here’s the stack we design for clients, from the ground up.

Layer 1 — Two connections that can’t fail together

The foundation is a business-grade fixed connection (NBN business fibre or Enterprise Ethernet) paired with a 4G/5G service on a different carrier. Different is the operative word, and it’s where most “redundant” setups quietly fail. Many providers resell the same underlying infrastructure: Boost and Belong both run on Telstra’s network, and plenty of budget brands ride Optus or Vodafone underneath. If your primary service is Telstra and your backup SIM secretly rides Telstra too, Wednesday morning takes them both. Ask every provider, in writing, which carrier network your service physically uses.

The same logic applies physically. Two fixed services entering the building through the same lead-in conduit share the same backhoe. For most businesses with 20–200 staff, pairing fixed with wireless is the cheapest way to get genuine diversity — different carrier, different technology, different path.

Layer 2 — Failover that happens without you

A backup link that requires someone to drive to the office and swap cables isn’t redundancy — it’s a scavenger hunt at 5am. A dual-WAN firewall or SD-WAN setup should detect the failure and switch traffic in seconds. Three details separate a setup that works from one that only looks like it does. First, health checks must test real internet reachability (pinging external targets), not just whether the cable has link — Wednesday’s failures were upstream, where a “link up” light lies to you. Second, plan for what rides the connection: VPNs and cloud phone systems need to re-register cleanly, and the backup link should prioritise critical traffic like payments and voice over someone’s YouTube tab. Third, decide failback behaviour so the network doesn’t flap between links while the carrier is still unstable.

One caveat: if you host anything on-premises that the outside world connects to, your static IP won’t follow you onto the backup link. Plan low-TTL DNS for failover, or better, move inbound services to the cloud where carrier outages can’t reach them.

Layer 3 — Phones that reroute themselves

Cloud phone systems like Teams Phone or hosted VoIP inherit your internet redundancy automatically — if the data link fails over, your numbers keep ringing. But configure the layer above it too: unreachable-call routing should divert your main number to nominated mobiles the moment the system can’t be reached, without anyone logging into a portal mid-crisis. And split those nominated mobiles across carriers. On Wednesday, teams whose phones were all Telstra had no plan B in their pockets — with dual SIM and eSIM, putting key staff on a second network costs almost nothing. Round it out with one out-of-band channel the whole team knows to check (a group chat that works on any network) and a printed contact tree for the day nothing digital cooperates.

Layer 4 — Payments, alarms and the systems you forgot

Dead EFTPOS terminals were Wednesday’s most visible business casualty. Terminals connected via Wi-Fi behind your failover-capable router inherit your redundancy for free; SIM-based terminals should be dual-SIM or multi-carrier where your bank supports it. Either way, know your bank’s offline transaction rules and have a manual fallback policy your counter staff can actually follow.

Then audit the dependencies you’ve forgotten about, because they’re almost always single-carrier: back-to-base alarms, duress buttons, lift phones, medical and on-call paging, cold-chain and environmental monitoring, courier apps, and the 4G modules inside other equipment. For aged care and healthcare providers, these aren’t conveniences — they’re the systems the strengthened Quality Standards assume will work.

Layer 5 — Power, people and a one-page plan

Failover is worthless if the comms cabinet is dark, and power and network events like to travel together. Put a UPS on the router, modem, core switch and Wi-Fi — sized for at least 30 to 60 minutes. Even Telstra’s own outage guidance recommends a UPS to keep your modem online.

Finally, the human layer: a one-page outage runbook that names who declares an incident, how the team communicates, the customer notice template, when to switch to manual processes, and who calls your IT partner. Then test it. A scheduled failover test each quarter — pull the primary link after hours and watch what actually happens — is the only way to know your redundancy is real. Untested failover fails at 4:30am, which is precisely when the network does. Build in the post-outage steps too: restart stubborn devices, reconcile offline payments, and remind staff that “Telstra compensation” calls that arrive out of the blue are scams.

Redundancy theatre vs the real thing

Plenty of businesses that went dark on Wednesday believed they had a backup. Here’s the difference between a setup that looks redundant on an invoice and one that keeps trading through a carrier failure.

For regulated businesses, availability is now a compliance question

If you operate in a regulated sector, Wednesday wasn’t just an operational headache — it was an audit question waiting to be asked. APRA-regulated entities are now working under CPS 230, which expects tolerance levels for disruption to critical operations and active management of service provider risk; your telco is a service provider in every sense that matters. ASIC-licensed advice firms carry an obligation to maintain adequate risk management systems, and five hours of being unreachable during market hours is a live test of them. Aged care providers under the strengthened Quality Standards need continuity of care systems that don’t stop working because a carrier’s timing nodes did. It’s worth noting the Essential Eight — vital as it is — is a security framework: it hardens you against attackers, not against your carrier having a bad morning. Availability needs its own plan.

The ACMA investigation may eventually produce findings, and Telstra says it will work with affected customers once services are restored. But as Optus customers learned in 2023, goodwill gestures rarely cover what a day offline actually costs — and no rebate restores a missed emergency call. Regulators are asking the carriers hard questions this week. Your clients, auditors and board will eventually ask you the same one: what happens to your business when the carrier goes down?

Frequently asked questions

Wednesday’s outage will fade from the headlines within a week. The architecture decisions it exposed won’t fix themselves. Whether your business shrugged it off or spent five hours taking cash and apologising down a mobile that wouldn’t connect came down to choices someone made — or didn’t make — long before 4:30am. As part of our managed IT services, we design, build and test exactly this kind of resilience for Australian businesses every day.