ISSB (IFRS S1 / S2) sustainability-related financial disclosure report outline

Structure (10 sections)

1. 1. Basis of preparation, reporting entity and timing

   Settle four things in the opening: (1) which standards the report is prepared under (for example "in accordance with IFRS S1 and IFRS S2", and if only partially applied, the extent of application); (2) whether the reporting entity is identical to the consolidation scope of the financial statements — if not, explain the difference and why; (3) whether the reporting period matches the financial reporting period; and (4) whether this report is published at the same time as the annual financial report or later. These four points determine whether investors can align sustainability information with financial data; leaving them vague draws a first-round query.

   Evidence to attach

   • · Consolidation scope reconciliation (financial consolidation vs sustainability reporting boundary, entity by entity with inclusion/exclusion reasons)
   • · Statement on alignment of the reporting period with the financial reporting period
   • · Basis-of-preparation statement and list of standards applied

   Common pitfalls

   • · Excluding overseas subsidiaries or compute facilities from the reporting boundary without saying so — investors spot the gap the moment they compare with the financial notes
   • · Claiming the report is "prepared with reference to ISSB" without stating which requirements were not applied; in an assurance engagement such vague wording is generally treated as a non-compliant statement
   • · Publishing the sustainability report months after the annual report without explaining why or how the two connect

2. 2. Governance: oversight body, management responsibilities and competence

   Set out who is accountable for sustainability-related risks and opportunities and how. Be specific: the name of the board or its committee and the clause of the terms of reference that grants the mandate; how often that body receives sustainability reporting and how many meetings actually took place in the period; how the board satisfies itself that it has, or can access, the right skills (for an AI company, compute electricity and model energy use are areas where boards commonly lack judgement — describing the external adviser arrangement is more credible than asserting competence); and whether sustainability targets feed executive remuneration, with weightings and metrics if they do. Do not write a prose version of the org chart — reviewers look for evidence that the mechanism actually ran during the period.

   Evidence to attach

   • · Board or sustainability committee terms of reference, with the sustainability mandate clause marked
   • · Minutes and agendas of board/committee meetings that addressed sustainability topics during the period
   • · Board resolution or remuneration policy extract linking executive pay to sustainability metrics
   • · Record of the board competence assessment and any external adviser engagement letters

   Common pitfalls

   • · Writing only that "the board attaches great importance" with no clause reference and no meeting count
   • · Parking ESG accountability on a part-time role with no budget or headcount, then being unable to describe its reporting line under questioning
   • · Claiming a remuneration link without disclosing weighting, metric or outcome

3. 3. Strategy: identified sustainability-related risks and opportunities and their time horizons

   List each sustainability-related risk and opportunity that could reasonably be expected to affect the entity’s prospects, and for each state whether it is a risk or an opportunity, whether the effect falls in the short, medium or long term, and how you define those three horizons (including how they relate to your strategic planning cycle and capital allocation cycle). Address head-on the risks an AI company is always asked about: compute electricity supply and price volatility; the carbon intensity of the grid where data centres sit and local constraints on power access; geopolitical disruption in the chip and hardware supply chain; water use and cooling for model training; and the very concrete revenue risk of customers cancelling procurement because carbon data falls short. On the opportunity side, do not stop at "the green transition creates market space" — name the product line and the customer segment.

   Evidence to attach

   • · Risk and opportunity register (identification date, owning function, basis for the horizon assessment)
   • · Compute and data-centre electricity and water ledgers with the carbon intensity of the relevant grids
   • · Supply chain geographic map and list of critical single-source dependencies
   • · Records of customers or tenders won or lost on sustainability grounds (redacted)

   Common pitfalls

   • · Naming "short, medium and long term" without defining them in years, so the metrics and targets section cannot line up
   • · Copying a generic industry risk list that shows no connection to your own compute footprint or customer mix
   • · Disclosing risks but not opportunities, or the reverse — both sides are required

4. 4. Strategy: effects on the business model, value chain, financial position and performance

   This section answers "what did these risks and opportunities actually change". Write it in three layers: (1) which parts of the business model and value chain are affected (own data centre vs rented cloud, in-house silicon vs purchased, choice of training location); (2) effects already reflected in current financial position and performance (for example the gross-margin impact of higher electricity prices, or new renewable power purchase costs incurred to meet customer carbon requirements); and (3) anticipated financial effects — where quantitative information is not yet available, qualitative information with an explanation is permitted, but the section must not simply be left blank. The practical advice for a Chinese AI company: it is far better to write "the current-period effect is concentrated in compute procurement cost, not yet separately captured, and will be presented as a separate account line next year" than to assert "not material" with no basis.

   Evidence to attach

   • · Financial workpapers for the affected cost lines (electricity, renewable power or certificate purchases, cooling, colocation)
   • · List of sustainability-related items in the capital expenditure plan with approval records
   • · Assumptions and sensitivity analysis behind anticipated financial effects (or a written explanation of why they cannot be quantified)

   Common pitfalls

   • · Dismissing the section with "no material financial effect" and offering neither basis nor follow-up plan
   • · Booking renewable power purchases under marketing expense while publicising them as an ESG achievement, so the financial and disclosure views do not reconcile
   • · Anticipated financial effects that contradict the capex plan (the disclosure says long-term risk is high while the budget allocates nothing)

5. 5. Climate resilience and scenario analysis

   Explain how you assess the resilience of your strategy and business model to climate-related changes, and disclose the scenario analysis method: which scenarios, their source, the time horizons, which operating locations and value-chain stages were covered, and when the analysis was performed. For a compute-intensive business, physical risk should come down to data-centre siting (cooling load and thermal throttling under extreme heat, water stress, flooding) and transition risk to electricity price and carbon cost pass-through plus the shift in customer and investor preference towards low-carbon compute. Where this is a first attempt and capability is limited, an approach commensurate with your circumstances is permitted — but the method and its limitations must be stated; silence is not an option.

   Evidence to attach

   • · Scenario analysis method note (scenario source, parameters, horizons, coverage, date performed)
   • · Physical risk screening results for data centres and principal offices
   • · Electricity price and carbon price sensitivity calculations
   • · Record of the scenario analysis conclusions being presented to the board

   Common pitfalls

   • · Stating only that "scenario analysis was carried out" without the scenario source or horizons, leaving readers unable to judge the conclusion
   • · Covering only owned offices while excluding the third-party data centres that account for the bulk of energy use
   • · Treating scenario analysis as a one-off consulting deliverable that is never updated, so the same paragraph is reused year after year

6. 6. Risk management: identification, assessment, prioritisation and monitoring processes

   Disclose the processes used to identify, assess, prioritise and monitor sustainability-related risks, and how those processes are integrated into overall enterprise risk management. Be explicit about the inputs (internal ledgers, supplier data, customer questionnaires, regulatory scanning), the assessment method and parameters, how priority is determined (the basis for likelihood and magnitude), and the frequency and ownership of the process. An AI company should explicitly connect its existing compliance and security risk processes (algorithm filing, model evaluation, data-export assessment) to the sustainability risk process, so reviewers do not read sustainability risk management as a separate paper exercise that nobody performs.

   Evidence to attach

   • · Risk management policy text with the clause incorporating sustainability risk marked
   • · Risk assessment workpapers and scoring sheets for the period
   • · Version history of the risk register (evidencing periodic update rather than one-off preparation)
   • · Mapping between sustainability risks and existing compliance risk processes

   Common pitfalls

   • · Risks in the risk management section that do not match those in the strategy section (five in the strategy, two in the register)
   • · A complete-sounding process with no workpapers evidencing that it ran during the period
   • · Asserting "integration into enterprise risk management" as a slogan while being unable to say where sustainability risks sit in the corporate risk matrix

7. 7. Metrics: greenhouse gas emissions and cross-industry metric categories

   Disclose absolute gross emissions for Scope 1, Scope 2 and Scope 3 separately, and state the accounting basis used. For Scope 2, say whether the figure is location-based or market-based (and label each where both are disclosed). For Scope 3, disclose which of the 15 categories are included and explain any exclusions — for an AI company the three largest are usually Category 1 (purchased goods and services, including cloud and compute procurement), Category 2 (capital goods, including GPU servers) and Category 11 (use of sold products, including end-user inference energy); do not drop them silently because "the data is not available". Also disclose exposure to transition and physical risk, capital deployed towards climate opportunities, and whether an internal carbon price is used and how it affects decisions. Version note: the ISSB issued targeted amendments to the IFRS S2 greenhouse gas disclosures in December 2025, effective for annual reporting periods beginning on or after 1 January 2027 with early application permitted — so a 2026 report is prepared under the unamended text by default; if you elect early application, say so in the basis of preparation and confirm whether your jurisdiction’s adopted version incorporates the amendments.

   Evidence to attach

   • · GHG inventory workpapers (see tmpl-ghg-inventory-workpaper)
   • · Dual location-based and market-based Scope 2 calculations with contractual instrument evidence
   • · Scope 3 category inclusion/exclusion decision sheet with the materiality basis
   • · Internal carbon price policy and a case where it actually affected an investment decision
   • · List of emission factors used, with source and data year

   Common pitfalls

   • · Giving a single Scope 2 figure without saying whether it is location-based or market-based — the single most queried item
   • · Reporting one aggregate Scope 3 number with no category breakdown and no statement of exclusions
   • · Using emission factors that are three years old without labelling the data year, so the figure cannot be recalculated
   • · Netting renewable certificates against Scope 1, or double-counting renewable power across the market-based and location-based figures

8. 8. Industry-based metrics and SASB reference

   State the sources you considered in identifying industry-relevant metrics. IFRS S1 requires consideration of other sources of guidance where no specific standard applies, IFRS S2 is accompanied by industry-based guidance, and the SASB Standards serve as the industry-metric reference within the IFRS framework. For the three industry classifications most relevant to Chinese AI companies — software and IT services, semiconductors, and hardware — the focus typically falls on data-centre energy and water, data security and customer privacy, supply chain management, and access to critical materials. Disclose metric by metric: which industry metrics you adopted, which you did not and why, and the definition and calculation basis for each. Do not pad the section with inapplicable industry metrics; that simply shows no industry matching judgement was made.

   Evidence to attach

   • · Industry classification determination (which SASB industry and why)
   • · Adopted / not-adopted industry metric table with reasons
   • · Calculation basis and underlying data source for each industry metric

   Common pitfalls

   • · Applying the software and IT services metrics when the business is really hardware and robotics manufacturing, so the metrics do not fit the business
   • · Reporting industry metric values without definitions or denominators, making them non-comparable year to year
   • · Omitting certain industry metrics without explanation, leaving readers to assume the numbers were unflattering

9. 9. Targets: setting, progress and restatement

   Each target needs six elements: the metric definition, the base year and base value, the target year, the target value (absolute or intensity), the coverage (which scopes, which entities), and whether the target is voluntary or required by law or listing rules. Also disclose progress in the period and the basis for that assessment. If carbon offsets are used to meet the target, disclose the figures before and after offsets separately, together with the type of offset used. If the base year or methodology changes (an acquisition is consolidated, an emission factor database is replaced), recalculate under your restatement policy and disclose the reason and the quantified effect — silently changing the basis is among the easiest ways to be assessed as a material misstatement in an assurance engagement.

   Evidence to attach

   • · Board or management approval record for each target
   • · Base-year recalculation policy and the recalculation workpapers for the period
   • · Target progress calculations by scope and by entity
   • · Where offsets are used: project documentation, retirement records, and before/after comparisons

   Common pitfalls

   • · Announcing "net zero by 2050" with no interim target and no annual progress basis, leaving investors nothing to verify
   • · Vague target coverage (silent on whether Scope 3 is included), so progress figures are not comparable
   • · Expanding the boundary through acquisition without recalculating the base year, producing an apparent "natural" decline in emissions

10. 10. Significant judgements, measurement uncertainty and prior-period errors

    Bring three things together here: (1) the significant judgements made in preparation (materiality screening, value-chain boundary, choice of estimation method); (2) amounts subject to high measurement uncertainty, with the sources of that uncertainty and sensitivity (for an AI company the inference-energy estimate in Scope 3 Category 11 is usually the most uncertain item); and (3) errors in prior-period reports and how they are corrected here. This is the section where professional readers judge the credibility of the whole report — disclosing uncertainty raises credibility rather than lowering it, while concealing it gets magnified in assurance.

    Evidence to attach

    • · List of significant judgements with the decision records (including rejected alternatives)
    • · Quantified or qualitative uncertainty analysis with sensitivities
    • · Prior-period error identification, quantified effect and correction note
    • · Approval record for any change in estimation method

    Common pitfalls

    • · A report in which the words "estimate" and "uncertainty" never appear, reading like marketing material
    • · Silently amending prior-period figures instead of disclosing an error correction
    • · Recording only the conclusion of a significant judgement, with no basis that can be reviewed