Sustainability report outline for issuers on the Shanghai, Shenzhen and Beijing exchanges (aligned to the MOF Basic Standard)

Structure (6 sections)

1. 1. General requirements and disclosure framework: governance, strategy, management of impacts, risks and opportunities, metrics and targets

   Guide No. 1 of the exchange preparation guides sets the general requirements and disclosure framework, and the MOF Basic Standard uses the same four elements — governance, strategy, management of risks and opportunities, metrics and targets — with information quality requirements in its Chapter 3. The opening should state which rules the report is prepared under (the exchange self-regulatory guide, the MOF Basic Standard, and whether the ISSB standards are also referenced), the reporting entity and period, the topic screening method and conclusion, and how the information quality principles (reliability, relevance, comparability, verifiability, timeliness and understandability) are applied. Note that the current Shanghai instrument is Self-Regulatory Guide No. 4 — Preparation of Sustainability Reports (January 2026 revision), effective 30 January 2026, which repealed the January 2025 version; confirm the version before drafting, and check the Shenzhen and Beijing equivalents against their own current texts.

   Evidence to attach

   • · List of the preparation basis with the version and document number of each rule, and its effective date
   • · Topic screening process record and conclusion
   • · Internal control note on meeting the information quality requirements
   • · Reporting entity and consolidation scope note

   Common pitfalls

   • · Using the chapter structure of the January 2025 guide without checking against the 2026 revision
   • · Presenting the topic screening conclusion with no process, leaving nothing to answer an exchange query with
   • · Claiming alignment with the ISSB while the metric definitions conflict with the domestic guide

2. 2. Addressing climate change: governance, risk identification and GHG accounting

   Guide No. 2 provides the preparation basis for the climate topic, including GHG accounting and climate risk identification. Cover: the governance arrangements for climate; the method and conclusions for identifying transition and physical risks; the greenhouse gas data (for Scope 1 and Scope 2, state clearly the data year and basis of the power emission factor used — China grid factors are published by the Ministry of Ecology and Environment by data year at national and regional level and must match where the electricity is consumed); and the mitigation measures and targets. For compute-intensive businesses, give data centres their own subsection setting out how energy is attributed across owned, colocated and public cloud compute and where the data comes from — domestic regulators and overseas customers ask the same question, so writing it once serves both.

   Evidence to attach

   • · GHG inventory workpapers (see tmpl-ghg-inventory-workpaper)
   • · Record of the source publication and data year of the power emission factor used
   • · Data-centre energy attribution note (owned, colocated and public cloud separately)
   • · Climate risk identification process record and conclusion
   • · Mitigation measure ledger and target progress calculations

   Common pitfalls

   • · Omitting the data year of the power emission factor, or applying the China national factor to every site worldwide
   • · Excluding public cloud compute energy entirely, leaving energy use visibly out of step with business scale
   • · Reduction targets that are inconsistent with the group’s overseas disclosures

3. 3. Pollutant discharge, energy use and water resource use

   The 2026 revision added Guides No. 3 (pollutant discharge), No. 4 (energy use) and No. 5 (water resource use), extending supervision from a climate-first approach to fuller environmental coverage — for AI companies with owned or colocated data centres, Guides No. 4 and No. 5 carry the most work. Energy use must be compiled by type (purchased electricity, self-generated for own use, fuels) with intensity metrics; water use must separate withdrawal source, consumption and discharge, with data-centre cooling make-up water the main item; pollutant discharge is usually light for pure software businesses but must be compiled on the environmental impact assessment and discharge permit basis where hardware or robot manufacturing (welding, coating, cleaning) is involved. Fold the data points of all three guides into the ledger structure of tmpl-ghg-inventory-workpaper in one pass rather than maintaining three separate sets of sheets.

   Evidence to attach

   • · Energy consumption ledger by type with intensity calculations
   • · Water withdrawal, consumption and discharge ledgers including data-centre cooling make-up water
   • · Discharge permits, EIA approvals and monitoring reports where hardware manufacturing is involved
   • · Consistency check against data filed with the environmental authorities

   Common pitfalls

   • · Reporting electricity only and omitting fuels and self-generated power, which does not match actual operations
   • · Counting office water only and omitting data-centre cooling systems
   • · Pollutant figures in the report that differ from those filed with the environmental authorities

4. 4. Social and governance topics: employees, supply chain, products and services, business conduct

   Write topic by topic on employee rights and occupational health, supply chain security and responsible sourcing, product and service quality, and business conduct (anti-corruption, fair competition, whistleblowing). Under products and services, an AI company should cite its existing compliance outputs in structured form: the security assessment and algorithm filing for generative AI services, the state of content labelling implementation, the conclusions of the personal information protection compliance audit, and minor-protection measures. The supply chain subsection should explicitly cover outsourced data labelling and content moderation — the labour node drawing increasing attention both at home and abroad. Note that this section is lifted directly by overseas customers and investors, so prepare the English version in parallel and keep the figures identical.

   Evidence to attach

   • · Workforce composition and occupational health and safety ledgers
   • · Supplier onboarding and assessment records, including labour checks on labelling and content moderation vendors
   • · Security assessment and algorithm filing evidence for generative AI services
   • · Content labelling implementation records and the personal information protection compliance audit conclusions
   • · Anti-corruption policy, training records and the whistleblowing case log

   Common pitfalls

   • · Writing only that "filing has been completed" without a filing number or a verifiable public reference
   • · A supply chain section that avoids the outsourced labelling workforce
   • · Chinese and English versions carrying different figures, which overseas investors find by cross-checking

5. 5. Metrics and targets, year-on-year comparability and restatement

   List each metric with its definition, compilation basis, denominator, data source and current-period value, together with the prior-period figure and an explanation of the movement. For targets, give the target value, base year, coverage, deadline and progress. Comparability is where domestic reports most often lose marks: changes in consolidation scope, adjustments to compilation basis and factor updates all need to be handled under a restatement rule and explained. Append a basis-change table at the end of the report setting out each definitional change against the prior period and its effect on the numbers — that table measurably reduces the number of exchange query rounds.

   Evidence to attach

   • · Metric definition and compilation basis note
   • · Current vs prior period comparison with variance analysis
   • · Basis-change table (what changed, why, and the effect on the figures)
   • · Target progress workpapers

   Common pitfalls

   • · Replacing figures after a basis change with no restatement note, so the trend chart misleads readers
   • · Giving absolute values without the denominator basis, making intensity metrics unverifiable
   • · Targets disclosed without a deadline or coverage, so progress cannot be judged

6. 6. One ledger, three views: reuse mapping for the domestic basis, the ISSB and EU customers

   Build a mapping table that takes each data point in the underlying ledger to its three uses: the exchange preparation guide data point, the ISSB data point (IFRS S1/S2, which is also HKEX Appendix C2 Part D), and the EU customer questionnaire or ESRS data point. Annotate every basis difference (location-based vs market-based Scope 2, the intensity denominator, the water accounting boundary) and keep a conversion note where they differ. The payoff is direct: one release of figures serves all three audiences, and any challenge points back to the same workpaper. Three parallel sets of sheets will eventually contradict each other, and being caught by cross-comparison costs far more than building the mapping table.

   Evidence to attach

   • · Three-way data point mapping (domestic guide / ISSB / ESRS or customer questionnaire)
   • · Basis differences and conversion notes
   • · Version control record for externally released figures
   • · Tie-out sheet across the publications on each basis

   Common pitfalls

   • · Three bases maintained by three teams, so the annual report, the ESG report and customer responses all carry different numbers
   • · A mapping table built once and never updated with the ledger, obsolete within six months
   • · Treating basis differences as technical detail not worth explaining, so a customer who spots one escalates it into doubt about the data as a whole