The SOCI Act imposes registration, mandatory incident reporting and a Critical Infrastructure Risk Management Program on responsible entities across eleven sectors — and, unusually, treats DATA STORAGE AND PROCESSING as a critical infrastructure asset class in its own right. It also gives government last-resort step-in powers to intervene directly in a serious incident, which is the most intrusive provision in this catalog.
Security of Critical Infrastructure Act
Statute Australia
The SOCI Act is the Australian regime most likely to apply to a technology company that does not think of itself as infrastructure, and the reason is a definitional choice made in the 2021–22 amendments: critical data storage or processing is its own asset class. A provider holding or processing business-critical data on behalf of a responsible entity in a covered sector can be regulated directly, rather than being managed as somebody else’s supply-chain risk.
- Eleven sectors - Communications, data storage and processing, financial services and markets, water, energy, health care and medical, higher education and research, food and grocery, transport, space technology, and defence industry.
- Register of Critical Infrastructure Assets - Ownership and operational control, reported to government.
- Mandatory cyber incident reporting - 12 hours for a critical incident having a significant impact; 72 hours for a relevant incident. The 12-hour clock is the shortest in this catalog.
- Critical Infrastructure Risk Management Program - Covering cyber, personnel, physical and supply chain hazards, with an annual board-approved report.
- Government assistance measures - Information gathering, action direction, and an intervention power allowing the Australian Signals Directorate to act directly on a system in a serious incident.
That last power deserves to be named plainly, because there is nothing else like it in the regimes I track. Every other cyber regime tells an operator what to do. SOCI contemplates the state doing it instead, on the operator’s systems, as a last resort. It was contested during passage and its practical use has been minimal — but it exists, and any risk assessment for an Australian critical-infrastructure asset has to account for it.
The 12-hour reporting clock is the operational constraint that shapes everything else. Twelve hours is not enough time to complete an investigation, which means it cannot be an investigation-complete report — it is a detection-and-notify obligation, and meeting it requires the detection, the triage path, the decision authority and the reporting channel all to be pre-built. A company that plans to work out who signs off during the incident will miss it.
The convergence point is worth noting for anyone already doing this work elsewhere. SOCI’s risk-management program, DORA’s ICT risk framework, NIS2’s supply-chain requirements and the DOJ bulk data rule’s vendor due diligence are four regimes asking variations of one question: who touches your critical systems and data, what is their security posture, and can you demonstrate you assessed it? Four jurisdictions, four legal theories, one control. An organisation that builds a genuine third-party risk capability once can answer all four; one that builds four compliance artefacts will maintain four and satisfy none of them well.
Where this applies
This is a sectoral regime: it binds the industries listed below, and companies outside them generally fall outside it.