Telstra was warned in January that critical timing hardware in its network needed to be replaced or updated, six months before the equipment failed and brought the carrier’s national network to a standstill, and it appears to be the only Australian carrier that failed to act.

Chief executive Vicki Brady and five senior executives are today facing an emergency Senate inquiry hearing to defend the telco’s handling of the 8 July outage that wiped mobile coverage for millions of Australians, cut payment systems, shut down train networks in NSW and Victoria and prevented more than 600 people from reaching Triple Zero.

The Senate Chairman described Telstra as “Incompetent” with the inquiry hearing that Technology Audit Partners have been bought as an independent organisation to review the failures of Telstra.

It’s also been revealed that there was no proper documentation to changes being made to the Telstra network resulting in staff not knowing what previous changes had been made to the design of their servers.

When asked how many other changes had not been documented Brady was unable to give any answers expect that issues are under investigation.

The inquiry heard tha at t its peak, the fault affected approximately 45 per cent of all calls and data sessions on Telstra’s mobile network, Brady has revealed, the company’s first public disclosure of the true scale of the failure after earlier estimates of “tens of thousands of services”.

The Warning Telstra Sat On

The Australian Financial Review has revealed that US manufacturer Microchip Technology notified carriers in January that two of Telstra’s synchronisation supply units, including the clock-like device that keeps time across the network, required replacement or updating ahead of a looming GPS “week rollover event”.

“The GPS system has a week rollover event once every 1024 weeks which relates to 19.69 years,” the notice said. For specific model devices made after 2008, the internal week rollover would occur between December 2025 and January 2027.

Microchip spelled out exactly what would happen if the warning was ignored. “When the internal week number rollover occurs, the GPS or GNSS engine will roll back in time 1024 weeks,” the company wrote, adding that “the system will remain locked to GPS and there will be no alarms to indicate the date/time roll back.”

ChannelNews understands that all Australian networks received the alert, with Telstra the only carrier to ignore it. A TPG spokesman told the AFR the affected part was replaced in 2024, while an Optus spokeswoman said the carrier “reviewed the advice, assessed our network and worked closely with the vendor” to resolve the issue in April.

Discontinued Hardware

Microchip manufactures the SyncServer S300, a key part of Telstra’s node management network, and it appears these end-of-life servers were never updated. The S300, which began life under Symmetricom before passing to Microsemi and then Microchip, has been discontinued and replaced by the S600/S650 and TP4100/TP4500 ranges, positioned as high-performance, enhanced-security enterprise-class GPS network products.

The time server used by Telstra is an internal hardware clock linked by GPS/GNSS to atomic clocks aboard satellites, serving correct time out to network nodes. It carries four dedicated and isolated Ethernet ports, one of them Gigabit Ethernet, sized to handle thousands of NTP requests per second while maintaining microsecond-calibre timestamp accuracy.

How the Network Fell Over

In its 25-page submission published on Friday, Telstra confirmed the failure began with exactly the kind of time rollback Microchip had predicted.

“On Wednesday 8 July 2026, following completion of maintenance work on a piece of equipment used for timing and synchronisation in our network, we restarted that equipment and this triggered an underlying software configuration and reset the date in that equipment to 2006,” the submission said.

“Over the next few hours, the incorrect date rippled slowly across the network, causing authentication certificates for software in other servers to become invalid. We have let our customers and Australians down, and for that we are deeply sorry.”

The submission noted an “intentional design change” to the unnamed piece of equipment that was not properly documented, and confirmed the device had not received a crucial software update. When asked directly about the Microchip notice, Telstra referred back to the submission rather than answering.

Brady Fronts the Senate

Opening proceedings at the inquiry, Brady told senators that “Telstra let Australians down” and that she is “deeply sorry” for a failure that struck at people’s ability to run businesses, stay in touch and reach emergency help.

Brady, who cut short a family holiday and flew into Sydney two days after the outage began, is walking the committee through a detailed timeline. A Network Time Protocol server at a Telstra site restarted with the wrong date after maintenance at 3.38am on the Wednesday, with call and data failures first detected at 4.20am as the fault spread through the mobile network.

She told the committee the company’s investigation is continuing and committed to handing the findings of an external expert review to the inquiry once complete.

The hearing is being run by the Senate committee originally formed to investigate the catastrophic Optus outage of September last year, which was linked to two deaths. The Telstra failure marks the second major Triple Zero outage in the past 12 months, sharpening scrutiny of the resilience of the nation’s emergency call system.

UPDATED:

Telstra has admitted that the network failure that crippled its services originated on a server that was 15 years old, with the carrier conceding it had received warnings about the hardware for years, including one just six months before the failure, and failed to act.

Chief executive Vicki Brady told a Senate inquiry that the failed device was an SSU 2000 timing unit manufactured by Symmetricom, now part of Microchip, and locally supported by Scientific Devices, after Greens senator Sarah Hanson-Young pressed her on the equipment’s identity and history.

The revelations have raised serious questions about why equipment at the heart of Telstra’s network was not operationally up to scratch, with the inquiry hearing the carrier had a prior warning about possible hardware problems in 2022, followed by another warning in January this year.

“You’ve had this for 15 years,” Senator Hanson-Young told the hearing.

Undocumented Design Change At Centre Of Failure

Brady conceded that a critical design change made to the Melbourne-based server was never documented, a failure that directly contributed to the outage when maintenance crews later worked on the machine.

“This particular server that operates in our network, there are three of them that do the same job. They’re all in different parts of the country. This one is based in Melbourne, and prior to this issue happening, there had been a fault on this server, and a design change was implemented,” Brady said.

“So it was working different to our standard design for how this server would operate in our network. That was not documented.”

According to Brady, when the maintenance team came to work on the server, they were required to power it down and take it offline to replace the chassis, the case the server sits in, because its backup power supply was not working.

“So they put in a brand new chassis. Once they repowered the server that sits within that, this particular GPS card behaved in a different way to what they expected,” she said.

Software Update Considered, Then Shelved

The inquiry heard that a reminder to install a software update on the server was issued in January this year, with evidence suggesting the update had been outstanding for several years.

Telstra claimed its teams did not simply ignore the update. The carrier said the update affected a feature that was not in use at the time, based on the way the server and its GPS card were originally configured.

However, the subsequent design change made to address a separate fault meant the update became relevant, and Telstra had neither documented the change nor implemented the software fix by the time of the failure.

Senator Hanson-Young asked Brady why there was not more official documentation around the outage, suggesting the gaps raised questions over Telstra’s “incompetence”.

Brady responded: “Firstly, the documentation that we have found that was not up to date was not related to the software update. That was related to a prior design change made on this particular server.”

Emergency Roaming Claims Under Scrutiny

Brady insisted that all affected Telstra customers who attempted to call Triple Zero during the outage but could not connect through the Telstra network successfully “camped on”, also known as emergency roaming, to alternative networks to complete their calls.

That claim is now under scrutiny, given a statistic Telstra itself submitted to the committee showing that only 172 of the 604 Triple Zero calls made during the outage successfully camped onto the Optus or Vodafone networks.

It remains unclear whether the remaining 432 calls failed to camp on, or whether those callers were in locations where Telstra was the only available network, a distinction that goes to the heart of whether the emergency roaming safety net actually worked.