Summary

  • The ACMA's investigation report on the 8 November 2023 nationwide Optus outage finds 2,145 unsuccessful emergency calls — 2,144 to 000 or 112, and one, by an Optus Networks end-user, to 106 — and books them by legal entity, not by brand: Optus Mobile Pty Limited carries 2,091 under section 15 of the Determination, Optus Networks Pty Limited 41 under section 15 (42 under section 19(1)), and Optus Internet Pty Limited 12 under both headings, while Optus Fixed Infrastructure Pty Limited is attributed none.
  • Of the 369 welfare checks required under section 28(1), 361 sit with Optus Mobile and 8 with Optus Networks, none with Optus Internet; only 183 welfare checks were conducted, section 28(2) exceptions covered 1,593 calls, and 28 employee test calls were excluded from the findings.
  • The wider statutory ledger runs far beyond the emergency calls: section 148(1) contraventions of 4,560 for Optus Mobile, 113 for Optus Networks, 24 for Optus Internet and 1 for Optus Fixed Infrastructure, and section 101(1) service-provider rule contraventions of 4,543, 92 and 24 respectively.
  • In November 2024, subsidiaries of SingTel Optus Pty Limited paid penalties totalling more than AUD 12 million — announced at group level, with no published split of the money against the entity-level fault ledger.
  • Fault, in this record, is individually documented and collectively priced. That gap is where durability of the repair is decided, and it is not yet proven.

The regulatory document that closes the book on the 8 November 2023 Optus outage is a number-heavy, partly redacted investigation report. It is easy to summarise badly. Most reporting on the enforcement outcome condensed it into one figure — a penalty of more than AUD 12 million paid by "subsidiaries of SingTel Optus Pty Limited" — and one image, a national carrier whose network failed for a day. Both are true. Neither is the interesting part.

The interesting part is that the report is built as a ledger of named corporate licensees, each with its own Australian Company Number, each booked against its own statutory obligations, and each with sharply unequal counts. Read at that resolution, the outage is not one story about one brand. It is four distinct compliance records, and they do not point in the same direction.

Who actually failed, and how much

The central finding is a total: 2,145 unsuccessful emergency calls during the 8 November 2023 outage. Of those, 2,144 were attempts to reach 000 or 112. One — by an Optus Networks end-user — was an attempt to reach 106, the text-based emergency relay for people who are Deaf, hard of hearing or speech-impaired. The report sets the full breakdown out at Attachment B (https://www.acma.gov.au/sites/default/files/2024-11/Investigation%20report%20-%20Optus%20outage%201Nov23%20%28redacted%29.pdf).

What distinguishes this report from a generic outage narrative is that it then assigns those calls to entities:

  • Optus Mobile Pty Limited (ACN 054 365 696): 2,091 contraventions of section 15 of the Determination. More than 97 per cent of every failed emergency call in the record sits with the mobile licensee.
  • Optus Networks Pty Limited (ACN 008 570 330): 41 contraventions under section 15, and 42 under section 19(1). The one-call difference between the two statutory headings is visible in the report itself, and any accurate quotation of this entity's count has to carry its heading with it.
  • Optus Internet Pty Limited (ACN 083 164 532): 12 contraventions under section 15 and 12 under section 19(1).
  • Optus Fixed Infrastructure Pty Limited (ACN 092 450 783): no failed emergency calls attributed. The entity appears in the record for a single section 148(1) contravention, but it carries none of the 2,145.

This distribution matters because it is a fault-location instrument. A brand-level penalty treats the outage as a property of "Optus". The entity ledger tells you which licence, which network estate and which operational organisation actually broke. The answer is heavily concentrated: the mobile estate failed, and it failed for almost every caller who tried to reach help. A fixed-infrastructure entity with zero attributed emergency-call failures is part of the same corporate group but is not part of this failure record — a fact the pooled announcement cannot convey.

The concentration carries through to the human safeguarding layer. Under section 28(1) of the statutory scheme, 369 welfare checks were required — follow-up contact with people who had tried to call emergency services and failed, because a failed call may mean someone who needed help and never got it. The report attributes 361 of those required checks to Optus Mobile and 8 to Optus Networks. None are attributed to Optus Internet. The duty to check on people sat, almost entirely, where the duty to carry their calls sat.

The welfare-check accounting is also the part of the record where the outcome is least settled. Of the 369 required checks, the report shows 183 conducted. Section 28(2) exceptions applied to 1,593 calls — 31 under section 28(2)(b) and 1,562 under section 28(2)(c) — and 28 test calls placed by Optus employees were excluded from the findings altogether. What that leaves is a denominator problem the report does not resolve in its indexed text: how many real people who tried to call for help were followed up with, and what happened to the remainder.

A redacted regulator report can prove that the obligation existed and that many checks were not performed. It cannot show the reader what became of the people behind the calls.

The wider statutory ledger

The emergency-call failures are the most legible part of the record, but they are not the largest. The report also books:

  • Section 148(1) contraventions of the Telecommunications (Emergency Call Service) requirements: 4,560 for Optus Mobile, 113 for Optus Networks, 24 for Optus Internet, and 1 for Optus Fixed Infrastructure.
  • Section 101(1) service-provider rule contraventions: 4,543 for Optus Mobile, 92 for Optus Networks, and 24 for Optus Internet.

The pattern repeats with even sharper concentration. Optus Mobile holds more than 96 per cent of the section 148(1) count and more than 97 per cent of the section 101(1) count. Whatever went wrong operationally in November 2023, the compliance exposure it generated landed overwhelmingly on one licensee — and that matters for what happens next, because penalties, follow-on claims and any future enforcement attach to a specific ACN, not to a brand.

What the penalty stage collapsed

In November 2024, the enforcement outcome was announced: subsidiaries of SingTel Optus Pty Limited paid penalties totalling more than AUD 12 million following the ACMA's investigation into the outage that affected 2,145 emergency calls and 369 required welfare checks https://www.acma.gov.au/articles/2024-11/optus-pays-12-million-penalty-triple-zero-outage. The regulator's announcement frames the result at group level https://www.acma.gov.au/articles/2024-11/acma-concludes-optus-outage-investigation.

Notice the two resolutions running side by side. The fault trail is entity-scoped: four companies, four ACNs, counts booked under named statutory headings. The money trail is group-scoped: more than AUD 12 million, paid by "subsidiaries", with no published split telling the public whether Optus Mobile paid in proportion to its 2,091 failed calls and 4,560 contraventions, or whether the amount was settled some other way.

The honest reading is that this gap is a resolution mismatch rather than proof of concealment. A negotiated enforcement outcome is not obliged to reproduce the evidentiary granularity of the underlying findings, and nothing in the public record shows how the payment was allocated internally. But the mismatch has consequences that a reader should see clearly:

  • It removes the per-entity price signal. If a specific licensee's breaches had a publicly known price, that number would inform future operational decisions inside that entity, inform claimants, and inform any future regulator. A pooled figure prices the group, not the licensee.
  • **It lets the brand absorb the headline."Optus paid $12 million" is a sentence about a household name. "Optus Mobile contravened section 15 on 2,091 occasions" is a sentence about a specific licence that a board and an operational team own. The first is easier to publish and easier to forget.
  • It sets up the durability test without answering it. Repair, in a multi-entity carrier, lives in the estate of the entity that failed. Knowing that the failure lived in the mobile estate tells you exactly where the evidence of a durable fix would have to appear.

What this record can and cannot prove

A redacted, number-heavy regulator report proves counts and legal attribution. It does not prove mechanism. The report can establish that 2,145 calls failed, that they are attributed to four entities in a specific distribution, that 369 welfare checks were required and 183 conducted, that exceptions and exclusions changed the denominators in specific ways. It cannot establish, in its redacted form, the internal decision chain that let a planned change take the emergency call path down, which teams decided what, or whether the fixes announced afterwards hold under load.

That limit is worth stating plainly, because the temptation with a document like this is to over-read the prose between the tables. The numbers are the evidence. The gaps — the redactions, the unresolved welfare-check remainder, the unallocated penalty — are not evidence of wrongdoing by anyone; they are the parts of the record the public cannot see, and they define exactly what a durable accountability story would still need to show.

Where the ledger meets the next failure

The final reason the entity ledger matters is the one that outlasts this report. Prior BTW reporting has already named Optus Mobile Pty Limited as the defendant in a later Federal Court proceeding over a separate emergency-call failure on 18 September 2025. The significance of that, read against this ledger, is structural: the same licensee that carried 2,091 of the 2,145 failed calls in November 2023 is the entity that reappears in later enforcement exposure. Attribution at entity level is what makes recurrence legible.

A brand-level record — one penalty, one outage, one announcement — cannot tell you whether the organisation that failed is the organisation that was repaired.

The record as it stands shows attribution, not cure. There are no published per-entity remediation commitments, no post-repair performance metrics per licensee, and no per-entity accounting of the payment. The November 2024 payment is compatible with deep repair and with a settled liability; the numbers alone do not distinguish them. What would distinguish them is evidence that the mobile estate — the entity with 2,091 failed calls, 361 required welfare checks and more than 4,500 section 148(1) contraventions — now performs differently, observed over time, under a named licensee rather than behind a group announcement.