APRA CPS 230 cloud vendor compliance after July 2026
A CPS 230 playbook for Australian financial institutions and cloud vendors covering contracts, fourth-party risk, recovery tests and incident reporting.

From 1 July 2026, APRA’s CPS 230 transition for legacy material service provider contracts ended, and targeted amendments took effect. For material service providers in the cloud, APRA CPS 230 makes contract terms, fourth-party visibility and scenario testing a joint workstream with Australian financial institutions, supported by AWS Artifact compliance reports and automated disaster recovery orchestration.
- CPS 230 generally commenced on 1 July 2025. For legacy contracts, the transition ended at the next renewal or on 1 July 2026, whichever came first. Targeted amendments also took effect on 1 July 2026.
- The regulated entity must notify APRA as soon as possible and within 24 hours of a disruption to a critical operation outside its tolerance.
- Material service provider contracts must require disclosure of other material providers on which the vendor materially depends to deliver the service. The regulated entity’s policy must address fourth-party risks to critical operations.
Map CPS 230 obligations to cloud vendor evidence
Start with the critical operation and trace every service it needs. A vendor is material if the institution relies on it for a critical operation or if it creates material operational risk. APRA also expects institutions to classify core technology providers as material unless they can justify a different classification. Assess the supplier across contracts. A provider can become material through several arrangements taken together. An insurer’s claims platform, for example, might depend on a SaaS vendor for workflow, a cloud provider for hosting and a separate service for policyholder identity.
| CPS 230 obligation | Cloud or SaaS vendor responsibility to agree | AWS and Azure evidence or tooling | What the regulated entity should retain |
|---|---|---|---|
| Contractual access and audit | Provide audit access and support APRA’s access to service records, data and on-site visits. | AWS Artifact audit reports and Microsoft Service Trust Portal materials support assurance reviews. | Executed terms, clause mapping, named contacts and a procedure for exercising access. |
| Severe but plausible scenario testing | Participate in joint exercises and supply recovery and failback evidence. | AWS Elastic Disaster Recovery drills and Azure Site Recovery test failover for supported workloads. | Exercise scope, observed recovery time, reconciled data loss, minimum service achieved and remediation owners. |
| Fourth-party disclosure | Identify material dependencies, provide contractually required notices and explain which services are affected. | AWS sub-processor disclosures and Microsoft’s Online Services Subprocessor List are starting points. | Workload-specific dependency map, notice history, concentration assessment and unresolved gaps. |
| Escalation for APRA’s 24-hour disruption window | Agree an early alert trigger, a staffed escalation route and a schedule for updates. | AWS Health organisational view and Azure Service Health provide provider-side signals. | Incident chronology, tolerance assessment, actions taken, expected recovery and proof of notification. |
Test the critical operation and rehearse escalation
CPS 230 requires a systematic business continuity testing programme covering all critical operations, including an annual exercise. Set pass criteria against the institution’s three tolerances for each operation: maximum disruption time, maximum acceptable data loss and the minimum service level under alternative arrangements. For a payments exercise, record when the team can accept and settle transactions, which records need replaying and how much volume the fallback can handle. A running replica is only one checkpoint.
Use automated disaster recovery orchestration to rehearse restoration of applications and data. AWS Elastic Disaster Recovery can launch drill instances. Azure Site Recovery can run a test failover for a virtual machine or a recovery plan. For a managed SaaS service, ask the vendor to demonstrate recovery with your team and provide timestamped results. Include a scenario where a shared identity or network dependency fails across the primary and recovery environments. Examine whether apparently separate suppliers rely on the same infrastructure when assessing concentration, as discussed in why Australia leans on US cloud infrastructure.
In the exercise, have the vendor send a first alert with the start time, affected services and known customer impact. The institution should record when the critical operation moved outside tolerance and rehearse sending APRA the nature of the disruption, action taken, likely business impact and estimated return to normal. Build separate checks into the runbook for CPS 230’s 72-hour operational risk incident requirement, applicable ASIC breach reporting, and Australia’s 72-hour ransomware payment reporting rule. Where two APRA notification categories apply, CPG 230 directs entities to use the shorter window.
Decide what to buy, renegotiate or replace
Make approval depend on the signed contract. Check APRA access, audit rights, subcontractor liability and the ability to terminate all or part of the service. For superannuation trustees, confirm that termination is available if continuing the arrangement would conflict with the duty to act in beneficiaries’ best financial interests. The July amendments provide limited contractual relief for specified non-traditional providers, such as central banks, where terms are standardised or there is no formal agreement. Assess cloud arrangements against the applicable contract requirements and document any APRA exemption relied on.
Evaluate tooling against a live procurement case. Ask suppliers to demonstrate how AWS Artifact compliance reports or Microsoft Azure financial services compliance materials cover the specific services in scope. Record the report period, contract version and control owner. If evaluating enterprise third-party risk management software, require a demonstration that it can link one critical operation to several vendors and a shared fourth party, route a change notice to an accountable owner and export the evidence behind the material service provider register. CPS 230 requires that register to be submitted annually.
Price recovery and exit together. Compare bids in AUD using the same assumptions for replication, standby capacity, test environments, data export, migration support and parallel running. Banks can use payments, insurers claims processing and super funds fund administration as acceptance scenarios. Ask SaaS bidders to restore a representative dataset with permissions and audit history intact, then demonstrate export and reconciliation in a replacement environment. Require a dated remediation plan and a repeat exercise where a bidder misses an agreed tolerance. For offshore support or recovery, bring compliance into the design review before committing. CPS 230 requires prior APRA notification for material offshoring and notification when a significant change is proposed.
Take the quizWhat Kind of Holiday Spender Are You? Take the Black Friday QuizStart →- 1Gold Silver Price November: Futures Rise Near Yearly Lows

- 2U.S. Emergency Alert Test Nov. 17 Is Not a Power Outage

- 32027 NFL Mock Draft Puts Ohio State’s Jeremiah Smith at No. 1

- 4Michael Clarke Says Sandpapergate Scars Drew Steve Smith to New York

- 5GTA 6 New Zealand Time Trick Is Not Guaranteed to Work

Lifestyle · Food & Health4 Reasons to Eat Oatmeal, According to Nutrition Experts
Lifestyle7 Signs Your Social Battery Is Running Low (and Why You Keep Cancelling Plans)
Lifestyle · Food & HealthThanksgiving Food Safety: How Long Leftovers Last and the Safe Turkey Temperature
Lifestyle7 Quiet Habits of Emotionally Intelligent People, According to Psychologists















