Join our Newsletter — 33% off our NHI Course
Home› Guides› China Cybersecurity Law Guide: 2025 Amendment, CII Duties…
Guide Regulations & Frameworks

China Cybersecurity Law Guide: 2025 Amendment, CII Duties and Higher Penalties

PDF for newsletter subscribers · free to read online

← All guides
By NHI Mgmt Group, edited by Lalit Choda Updated 27 October 2026 15 min read
On this page

China's Cybersecurity Law (CSL) was amended for the first time on 28 October 2025, and from 1 January 2026 the penalties for security failures are far higher, tiered by consequence and aimed at responsible managers as well as companies. The CSL has applied since 1 June 2017 and remains the base law for network security in mainland China, sitting beside the Data Security Law and the Personal Information Protection Law. This guide explains who the law covers, the duties of network operators and critical information infrastructure (CII) operators, what the 2025 amendment changed, how enforcement works, what it means for machine identities and AI systems, and a practical compliance sequence. It is for security, legal and engineering leads at multinationals with operations, networks or customers in China. Status line: checked against the official amendment decision published by the Cyberspace Administration of China (CAC) and the 2016 text on 27 October 2026.

Key takeaways

  • Amended, not replaced: the NPC Standing Committee adopted the amendment decision on 28 October 2025, effective 1 January 2026, adding articles on Party leadership and artificial intelligence and rewriting the penalty chapter.
  • Fines now scale with consequences: serious consequences such as large-scale data leaks bring RMB 500,000 to 2 million, and especially serious consequences such as a CII losing its main functions bring RMB 2 million to 10 million for the entity.
  • Base fines no longer wait for a refusal: under the amended Article 61 a network operator that fails its security duties can be fined RMB 10,000 to 50,000 alongside the warning, where the 2016 text fined only after refusal to correct or harm.
  • Personal information and data export penalties move out: personal information violations and CII data export breaches are now punished under other laws, chiefly the PIPL and the Data Security Law framework, and Article 42 expressly requires PIPL compliance.
  • Incident reporting is measured in hours: the CAC's incident reporting measures, in force since 1 November 2025, require CII operators to report within one hour and other network operators within four hours.

Scope: who and what is covered

The CSL applies to the construction, operation, maintenance and use of networks within mainland China (Article 2). Its main duty holders are defined broadly.

Duty holderDefinition or testMain obligations
Network operatorNetwork owners, managers and network service providers (Article 76 of the 2016 text)Multi-level protection duties, incident plans and reporting, real-name verification for certain services, personal information security, cooperation with authorities
CII operatorOperator of infrastructure whose damage could seriously endanger national security or the public interest, in sectors such as energy, finance and transport (2016 Article 31); identified by sector protection departments under the 2021 CII RegulationsAll network operator duties plus security management body, background checks, training, backups, drills, annual assessment, procurement security review, data localisation
Network product and service providersSuppliers of network products and servicesMeet mandatory national standards, fix flaws, do not install malicious programs (2016 Article 22); critical network equipment and security products need certification or testing
Foreign organisations and individualsActivities outside China that harm China's cybersecurity (amended Article 77)Legal liability; in serious cases asset freezes and other sanctions

The amendment renumbered the law: two new articles (3 and 20) shift most later articles by two. This guide uses the post-2026 numbers where the official decision gives them and notes the 2016 number where it helps (for example, the multi-level protection duties were Article 21 and are referenced as Article 23 in the amended penalty article).

Where CII status comes from

The CSL does not set a numeric threshold for CII. Under the Critical Information Infrastructure Security Protection Regulations (State Council Decree No. 745, effective 1 September 2021), sector protection departments write identification rules, identify CII in their sector and notify operators (Articles 8 to 10). Operators must report significant changes, after which re-identification follows within three months (Article 11). If you have not been notified, you are a network operator, but sector regulators may still apply CII-style expectations.

Key obligations

Network operators

  • Multi-level protection duties (2016 Article 21): internal security management rules and named responsible persons; technical measures against viruses, attacks and intrusions; monitoring and logging of network operations and security incidents, with network logs kept for at least six months; data classification, backup of important data and encryption. The detailed scheme is the Multi-Level Protection Scheme; see the China MLPS 2.0 Guide.
  • Incident plans and reporting (2016 Article 25): written emergency response plans, prompt handling of vulnerabilities and attacks, and immediate activation of the plan and reporting to authorities when an incident occurs.
  • Real-name verification (2016 Article 24): operators providing network access, domain registration or certain messaging services must require users' real identity information.
  • Cooperation (2016 Article 28): technical support and assistance to public security and national security organs acting lawfully.
  • Personal information (amended Article 42 onward): handling must comply with the CSL, the Civil Code and the PIPL; breaches require remedial steps, notice to users and reports to authorities. See the China PIPL Guide.

CII operators (in addition)

DutySourcePractical evidence
Dedicated security management body and head; background checks for the head and key staff2016 Art. 34(1); CII Regs Arts 14-16Organisation chart, appointment records, screening records
Regular education, training and skills assessment2016 Art. 34(2)Training plan and completion records
Disaster recovery backups of important systems and databases2016 Art. 34(3)Backup schedules and restore tests
Incident response plans and regular drills2016 Art. 34(4)Drill reports
National security review for products and services that may affect national security2016 Art. 35; CII Regs Art. 19Procurement records and review outcomes
Security and confidentiality agreements with providers2016 Art. 36; CII Regs Art. 20Signed agreements
Store personal information and important data collected in China within China; security assessment for necessary outbound transfers2016 Art. 37 (amended Art. 39)Data maps, assessment filings
Security testing and risk assessment at least once a year, in-house or by a provider2016 Art. 38; CII Regs Art. 17Annual assessment report and remediation
Report major incidents and threats to the protection department and public securityCII Regs Art. 18Incident log with timestamps

Incident reporting deadlines

The CAC's Measures for the Administration of National Cybersecurity Incident Reporting took effect on 1 November 2025 and give the CSL's reporting duty concrete clocks. Incidents are graded in four levels (especially major, major, larger, general) using a classification guide annexed to the measures and based on GB/T 20986-2023.

Operator typeReport toDeadline
CII operatorsProtection department and public security authoritiesAs soon as possible, no later than 1 hour
Central and state organs and direct subordinate unitsTheir department's cybersecurity work bodyAs soon as possible, no later than 2 hours
All other network operatorsProvincial CACAs soon as possible, no later than 4 hours

Which incidents must be reported depends on the grading criteria in the annex; see the China Cybersecurity Incident Reporting Measures Guide for the thresholds and report contents.

Timeline and current status

Status as of 27 October 2026. The amendment decision is published on the CAC website; this summary relies on a translation of that Chinese text, and the Chinese version is authoritative.

DateEvent
November 2016CSL passed
1 June 2017CSL takes effect
1 September 2021CII Security Protection Regulations (Decree No. 745) take effect
28 October 2025NPC Standing Committee adopts the amendment decision (18th session of the 14th NPCSC)
1 November 2025National Cybersecurity Incident Reporting Measures take effect
1 January 2026Amended CSL takes effect

What the 2025 amendment changed

  • New Article 3: cybersecurity work follows Communist Party leadership and the overall national security concept, balancing development and security.
  • New Article 20 on AI: state support for basic AI research, algorithms, training data and computing infrastructure; AI ethics norms, risk monitoring and safety supervision; and use of AI to improve cybersecurity protection. It is a policy article, not a set of operator duties.
  • Article 42: a new paragraph requiring personal information handling to comply with the CSL, the Civil Code and the PIPL.
  • Penalty chapter: higher, consequence-tiered fines (next section); a new Article 63 for selling critical network equipment and security products without required certification or testing; Article 64 now expressly includes closing applications; a new Article 73 applying the Administrative Penalty Law's mitigation and exemption rules.
  • Article 71: personal information violations and CII export breaches are now handled under other laws and regulations rather than the CSL's own penalties.
  • Article 77: extraterritorial liability broadened from attacks on CII to activities that harm China's cybersecurity.

Enforcement and penalties

The CAC coordinates enforcement, with public security and sector regulators acting within their remits. The amended penalty structure, from the official decision:

ViolationEntity fine (RMB)Responsible individuals (RMB)2016 comparison
Network operator fails security duties (Arts 23, 27)10,000 to 50,000 with the warning; 50,000 to 500,000 if it refuses to correct or causes harm10,000 to 100,00010,000 to 100,000 only after refusal or harm; managers 5,000 to 50,000
CII operator fails CII duties (Arts 35, 36, 38, 40)50,000 to 100,000 with the warning; 100,000 to 1 million on refusal or harm10,000 to 100,000100,000 to 1 million only after refusal or harm
Serious consequences, e.g. large-scale data leak, partial loss of CII function500,000 to 2 million50,000 to 200,000No separate tier
Especially serious consequences, e.g. CII loses main functions2 million to 10 million200,000 to 1 millionNo separate tier
Critical network equipment or security products sold without certification or testing (Art. 63)20,000 to 100,000, or 1 to 5 times illegal gains of RMB 100,000 or more; suspension or licence revocation if seriousCheck the official textNew article
CII uses products or services that have not passed security review (Art. 67)1 to 10 times the procurement amount10,000 to 100,000Existing duty, renumbered
Prohibited content duties (Art. 69)50,000 to 500,000; 500,000 to 2 million if serious; 2 million to 10 million if especially seriousUp to 1 million at the top tierConsolidated and raised

Fines sit alongside warnings, rectification orders, suspension of business, closure of websites or applications, and licence revocation. Mitigation under the Administrative Penalty Law (Article 73) means prompt correction and cooperation can reduce or remove a penalty. The CII Regulations carry their own penalty provisions (Articles 39 to 42), for example for failing to report major incidents.

What it means for AI agents and non-human identities

The CSL does not mention service accounts, API keys or AI agents, but its duties are written around networks and data, and in practice they are met or broken through machine identities.

CSL dutyMachine identity or AI implicationControl
Logging and six-month log retentionActivity by integrations and agents must be attributable, not hidden behind shared accountsOne identity per workload and agent; logs that record which identity acted, kept at least six months
Technical measures against intrusionLeaked keys and long-lived tokens are a common intrusion pathSecrets in a vault, rotation, short-lived credentials
Data localisation for CII (amended Art. 39)Global SaaS connectors and AI services can move important data or personal information out of China through API callsInventory of outbound integrations from China environments; egress controls on service accounts and agent tools
Security review of products and services for CIIAI platforms and identity tooling used in CII may be in scopeProcurement review records covering AI and identity services
One-hour and four-hour incident reportingScoping a compromised credential quickly decides whether you meet the clockCredential-to-system mapping maintained before an incident
Article 20 AI safety supervisionSignals regulatory attention to AI risk; detailed duties sit in AI-specific rulesTrack the generative AI measures and labelling rules; see the China Generative AI Measures and AI Labelling Rules Guide

For multinationals the practical risk is a global identity platform or AI service with standing access to China-hosted systems: it creates both a data export question and an intrusion path. Treat China environments as a separate trust zone with their own machine identities. See the Service Account Security Guide and the Secrets Management Guide.

Compliance steps

  1. Confirm your role: record whether each China entity is a network operator, whether any protection department has notified CII status, and whether you supply critical network equipment or security products.
  2. Complete MLPS grading: grade systems and implement the multi-level protection duties, including six-month log retention, classification, backup and encryption.
  3. Rebuild incident reporting: align runbooks to the one-hour (CII) or four-hour (others) clocks, the four-grade scheme and the 12387 reporting channels; name a China-based decision maker who can report without waiting for global approval.
  4. Map data flows and integrations: list every service account, API key and AI tool that moves data out of China environments; for CII, apply localisation and the outbound assessment route; for all, apply PIPL transfer rules.
  5. Review procurement: for CII, run security review where products may affect national security and sign security and confidentiality agreements; for suppliers, confirm certification or testing of critical equipment and security products.
  6. Brief responsible managers: individual fines now reach RMB 1 million at the top tier, so accountable executives need to know their duties.
  7. Assess annually: CII operators test and assess at least once a year; other operators should still assess to evidence the duties they certify under MLPS.
  8. Keep the record for mitigation: document prompt correction and cooperation, which matter under Article 73.

Decision guidance

  • Small foreign subsidiary, no CII notice: focus on MLPS, incident reporting within four hours, PIPL and transfer routes.
  • CII operator or supplier to CII: expect procurement security review, localisation, annual assessment and the one-hour clock; budget for China-local security operations.
  • Global platform teams: decide whether China runs on separate tenants and identities; shared global admin and integration accounts make every China duty harder to evidence.

Detection, measurement and framework mapping

  • Reporting timeliness: percentage of reportable incidents reported within one or four hours.
  • Log coverage: share of in-scope systems with logs that attribute actions to individual human or machine identities and are retained for six months or more.
  • Outbound integration inventory: number of service accounts and AI tools with cross-border data paths, and the share with a recorded transfer basis.
  • MLPS status: share of systems graded, filed and assessed on schedule.
  • Remediation age: open findings from the annual assessment past their due date.
CSL dutyNIST CSF 2.0NIST SP 800-53 Rev. 5
Responsible persons, security management bodyGV.RR-02PM-9
Logging, six-month retentionPR.PS-04AU-11
Backups of important dataPR.DS-11CP-9
Background checks, trainingPR.AT-01PS-3, AT-2
Procurement security review, provider agreementsGV.SC-05, GV.SC-06SA-4, SR-6
Incident plans, drills and reportingID.IM-04, RS.CO-02IR-8, IR-6
Annual testing and risk assessmentID.RA-01RA-3, RA-5
Identity for workloads and agentsPR.AA-01IA-4, AC-2

Common failure patterns

  • Relying on pre-2026 penalty assumptions: budgeting for fines that only applied after a refusal to correct, when the amended text fines alongside the first warning.
  • Global incident processes that cannot meet the clock: escalation through a head office in another time zone breaks a one-hour or four-hour deadline.
  • Shared global service accounts: one integration identity spans China and other regions, so logs cannot show what happened in China and revocation affects everything.
  • Invisible data export: SaaS connectors, monitoring agents and AI assistants send logs or records abroad without anyone treating them as transfers.
  • Using old article numbers: policies and contracts cite 2016 numbering that no longer matches the law after renumbering.
  • Waiting for CII notice: suppliers to CII discover procurement review requirements late in a sales cycle.

Practitioner checklist

  • Record each China entity's status: network operator, notified CII operator, or equipment supplier.
  • Update policies and contracts to the post-2026 article numbering.
  • Complete MLPS grading and keep logs for at least six months with per-identity attribution.
  • Rewrite incident runbooks for one-hour (CII) and four-hour reporting with a China-based decision maker.
  • Inventory service accounts, API keys and AI tools with cross-border data paths.
  • Run procurement security review and provider agreements for CII.
  • Brief responsible managers on individual fines up to RMB 1 million.
  • Run and document the annual security test and risk assessment.

Standards and references

Related NHI Mgmt Group resources: China PIPL Guide · China Data Security Law and Network Data Security Regulations Guide · China MLPS 2.0 Guide · China Cybersecurity Incident Reporting Measures Guide. This guide summarises China's Cybersecurity Law as amended with effect from 1 January 2026 and is not legal advice.

PDF for subscribers

Take this guide with you

Every section, table and checklist in one PDF, with links to the sources and related guides. Subscribe to our weekly NHI & AI Identity Journal and we’ll email you a download link.

Free weekly newsletter

Subscribe to the NHI & AI Identity Journal

The latest on NHI and Agentic AI security – articles, research, breaches, news and events every week.

Bonus 33% off our NHI Course when you subscribe.

NHIMG Editorial Note
Researched and written by the NHI Mgmt Group editorial team using AI tools, under the editorial direction of Lalit Choda. Facts are drawn from the sources listed on this page and checked by our automated quality process. Last updated 27 October 2026. Spotted an error? Use our contact page.
NHI Mgmt Group, the independent authority on Non-Human Identity, IAM and Agentic AI security. nhimg.org

PDF Downloads for Newsletter Subscribers

Get the guide

Enter your details and we’ll email you a download link.

The full guide is free to read online. The PDF is a thank you for subscribing.

Fill in all fields marked * and tick the PDF box to continue.