Most climate software pitched to Indian industry over the past decade was sold on goodwill. A sustainability team liked the dashboard, a pilot ran at one plant, and the budget line disappeared at the next cost review. That pattern is changing, not because executives have become more concerned about the climate, but because three sets of rules now attach legal and financial consequences to emissions data: SEBI's BRSR Core disclosure and assurance regime, the Carbon Credit Trading Scheme and its emission intensity targets, and the European Union's Carbon Border Adjustment Mechanism for exporters.
For anyone building climate tech AI in India, the useful question is no longer "who cares about emissions?" but "who is now obliged to produce a number, by when, to whom, and with what penalty if it is wrong?" Each rule creates a different buyer, a different data problem and a different tolerance for error. A model that estimates supplier emissions for a disclosure report is a different product, sold to a different person, from a model that forecasts whether a cement kiln will meet its intensity target by March.
This article maps those obligations onto product categories, for founders choosing a problem, corporate teams deciding what to buy and investors separating regulation-backed demand from interest that will fade. It is explicit about where the rules are still moving.
Three rules, three kinds of obligation
BRSR Core: disclosure that must survive an assurer
SEBI's circular of 12 July 2023 (SEBI/HO/CFD/CFD-SEC-2/P/CIR/2023/122) introduced BRSR Core, a subset of the Business Responsibility and Sustainability Report built around nine ESG attributes. KPMG India's note on the circular lists them: greenhouse gas footprint, water footprint, energy footprint, embracing circularity, employee wellbeing and safety, gender diversity, inclusive development, fairness with customers and suppliers, and openness of business. The GHG footprint attribute asks for total Scope 1 emissions, total Scope 2 emissions and Scope 1 plus 2 emission intensity. The energy footprint asks for total energy consumed, the share from renewable sources and energy intensity. The circular states that intensity ratios are to be based on revenue adjusted for purchasing power parity.
What turns this from a reporting exercise into a buying trigger is assurance. The circular sets a glide path for reasonable assurance of BRSR Core: the top 150 listed entities by market capitalisation in FY 2023-24, the top 250 in FY 2024-25, the top 500 in FY 2025-26 and the top 1,000 in FY 2026-27. Reasonable assurance means an external assurer must be able to trace a reported tonne of CO2 equivalent back to meter readings, fuel invoices and emission factors. A spreadsheet assembled by a consultant in the last fortnight of the financial year does not survive that.
The circular also introduced value chain disclosures for the top 250 listed entities, covering upstream and downstream partners that together make up 75% of purchases or sales by value, with stated scope, assumptions and estimates. SEBI subsequently deferred and eased the value chain requirements in December 2024 changes. A founder should treat value chain demand as real but slower, and should check the current SEBI position before building a sales plan on a specific date.
The Carbon Credit Trading Scheme: targets with a price attached
The Carbon Credit Trading Scheme, notified by the Ministry of Power on 28 June 2023 according to the International Energy Agency's policy database, has two parts. Under the compliance mechanism, obligated entities in energy-intensive sectors receive greenhouse gas emission intensity targets. Entities that beat their target earn carbon credit certificates; those that miss it must buy certificates through the Indian Carbon Market. Under the offset mechanism, non-obligated entities can register projects that reduce, remove or avoid emissions and earn certificates voluntarily. The Bureau of Energy Efficiency's carbon market pages publish detailed procedures for both mechanisms and run the accreditation of carbon verification agencies.
The targets are no longer hypothetical. The Ministry of Environment, Forest and Climate Change issued the Greenhouse Gases Emission Intensity Target Rules, 2025 on 8 October 2025, after a draft in April. ETV Bharat reported that the first notification covered 282 industrial units in aluminium, cement, pulp and paper and chlor-alkali, with a 2023-24 baseline and compliance years of 2025-26 and 2026-27. The reported reductions from baseline over the period were about 3.4% for cement, 5.8% for aluminium, 7.1% for pulp and paper and 7.5% for chlor-alkali, measured in tonnes of CO2 equivalent per tonne of product. On 13 January 2026 the government added 208 obligated entities in petroleum refining, petrochemicals, textiles and secondary aluminium, bringing the total to 490, as reported by The Statesman.
The penalty is what makes this a budget item. Entities that miss a target and do not buy enough certificates face environmental compensation of twice the average trading price of carbon credits in that compliance year, levied by the Central Pollution Control Board. The JSA law firm's analysis of the draft rules gives a worked example: one aluminium smelter's intensity moving from a baseline of 13.4927 to 13.2260 in 2025- 26 and 12.8259 in 2026-27. Differences in the second and third decimal place now have a rupee value.
CBAM the customer's regulator
The third rule is not Indian at all. The European Commission's CBAM pages state that the mechanism covers selected imports of cement, iron and steel, aluminium, fertilisers, electricity and hydrogen, and applies in its definitive regime from 1 January 2026. Importers must buy and surrender CBAM certificates matching the embedded emissions of the goods they import. For an Indian exporter of aluminium or steel, this means the European buyer will ask for installation-level embedded emissions data, and will either pay for verified data or price in a default value that is usually worse.
Why the three rules create different buyers
The rules overlap on some sectors, aluminium and cement most obviously, but they are owned by different people inside the same company. BRSR Core sits with the company secretary, the chief financial officer and the sustainability head, and the person who signs off is worried about the assurer. CCTS sits with plant heads, energy managers and operations leadership, and the person who signs off is worried about missing a target and paying twice the market price. CBAM sits with export sales and the customer relationship, and the person who signs off is worried about losing a European order.
This determines the sales motion, the integration depth and the tolerance for model error. A disclosure buyer wants an audit trail. An operations buyer wants a forecast that changes what the plant does next week, reconcilable with the control system. An export buyer wants a defensible product-level number in the customer's format.
A demand map for climate tech AI in India
The table below is the article's main contribution. It ties each obligation to the product category it creates, the internal buyer, the data the product must touch, and the strength of the regulatory pull. "Pull strength" is our classification, based on three tests: whether the obligation is legally binding now, whether there is a direct financial penalty or price, and whether an external party (assurer, verifier or customer) checks the output.
Three readings follow from the map.
First, the strongest pull is operational, not reporting. A disclosure tool helps a company report what happened. An intensity tool helps a plant change what happens, and under CCTS every tonne of intensity avoided is either a certificate earned or a penalty avoided. The money is closest to the process.
Second, the same plant data serves all three rules. Scope 1 emissions for BRSR, intensity against a CCTS target and embedded emissions for CBAM all start from the same fuel flows, power draws and production counts. A venture that owns clean, timestamped plant data has a path into all three buyers. A venture that starts from disclosure templates has to fight its way back to the plant.
Third, error tolerance falls as the checker becomes stricter. Supplier estimates for value chain disclosure can rely on spend-based factors; a CCTS compliance figure will be checked by an accredited verifier. Model claims should match the buyer's tolerance.
The industrial emissions base the rules sit on
The rules concentrate on industry because that is where the measurable, attributable emissions are. India's Fourth Biennial Update Report to the UNFCCC, as summarised by Down To Earth in January 2025, put total national emissions in 2020 at 2,437 million tonnes of CO2 equivalent including land use, with the energy sector the largest contributor at 2,238 million tonnes. Within energy, Down To Earth reports that energy industries accounted for 56% and manufacturing for 17%. Industrial processes and product use, which includes the process emissions from cement clinker and aluminium smelting, accounted for 8.06% of the total, or 239 million tonnes.
Two points matter for product builders. Industrial emissions are a mix of combustion (fuel burned for heat) and process chemistry (CO2 released by the reaction itself, as in clinker production). AI can optimise combustion efficiency and energy use directly; it can only influence process emissions through recipe and material choices such as clinker substitution. And the emissions are concentrated in a few hundred large sites, which is exactly the population the CCTS has started to name. A climate AI venture targeting heavy industry in India has a knowable customer list of a few hundred plants, not a long tail.
Which product categories are buildable now
Taking the map and the emissions base together, four categories look buildable for a venture today, in roughly descending order of regulatory pull.
Intensity forecasting and process optimisation for obligated plants. This is the category with the clearest money: a plant that misses its target pays twice the average certificate price. The product forecasts end-of-year intensity from current operations, attributes deviations to specific units, and recommends operating changes. It requires deep integration with process historians and control systems, and the sale is to operations, not sustainability. This category overlaps with industrial AI more broadly. PredCo, an industrial AI and predictive maintenance venture in the PanScience portfolio, works on manufacturing plant data of this kind, and the lesson from that adjacent field applies here: the hardest part is reliable, contextualised plant data, not the model.
Audit-grade emissions accounting. The BRSR Core glide path reaches the top 1,000 listed entities in FY 2026-27, so the number of companies needing assurance-ready Scope 1 and 2 data is growing every year. The product's value is evidence: every number linked to its source document and emission factor version. Competition here includes the large accounting and consulting firms and global carbon accounting platforms, so an Indian venture needs an advantage in integration with Indian utility bills, fuel records and ERP configurations.
MRV preparation for verification. As accredited carbon verification agencies begin work under the CCTS, there is a workflow problem between the plant and the verifier: calibration records, data gaps, recalculations. This is less glamorous than optimisation but has a clear, recurring buyer in each obligated entity.
CBAM embedded emissions for exporters. This is a narrower market, limited to covered goods exported to the EU, but the pull is direct because the customer asks for the data. The work overlaps heavily with CCTS data for aluminium and steel producers.
Supplier emissions estimation for value chain disclosure is a fifth category, but given the December 2024 easing and the estimate-tolerant nature of the requirement, it is a weaker standalone business than it looked in 2023.
PanScience lists an emissions AI venture among its stealth categories. We do not discuss it here, but its existence is consistent with the reading above: studios and founders are moving towards the parts of climate software where regulation has started to set deadlines and penalties.
What the rules do not yet create
It would be a mistake to read the map as a guarantee of demand. Three gaps remain.
The certificate market needs liquidity before credit position forecasting becomes a product. Until trading volumes are established, treasury teams will not pay for sophisticated optimisation of something that barely trades.
The offset mechanism depends on approved methodologies. Products built for offsets are only as valuable as the methodologies the Bureau of Energy Efficiency adopts, which makes this category exposed to rule changes in a way that the compliance mechanism is not.
Most of the rules regulate intensity, not absolute emissions. A plant can meet an intensity target while growing total emissions with output. That shapes the product: the metric to optimise is tonnes of CO2 equivalent per tonne of product, and a product that reports only absolute tonnes will miss what the plant head is measured on.
How corporates and founders should use the map
For a corporate, the map is a procurement tool. Before buying climate software, identify which of the three rules applies, who owns it internally, and which external party will check the output. Then buy for that checker. A tool that impresses the sustainability team but cannot produce what the verifier asks for is a cost without a return. Where the corporate is deciding whether to build such a capability itself, partner with a venture or buy a product, a separate PanScience Research piece covers build versus buy versus co-build for enterprise AI.
For a founder, the map is a positioning tool. Pick one obligation and one buyer, and own the plant data that serves it. Expand from there into the adjacent rules that use the same data. Early design partners matter disproportionately in this market because integration with a plant's control systems is slow and specific; a separate PanScience Research piece on lighthouse partners covers how to choose them.
For investors, the pull-strength column is a diligence shortcut. Ask a climate AI company which obligation its revenue depends on, whether that obligation is binding now, and who checks the output. A company whose answers sit in the high-pull rows is selling into a deadline. A company whose answers sit in the medium rows is selling into good intentions, which in Indian industry has historically meant pilots that do not convert.
Conclusion
Regulation has done for climate tech AI in India what years of awareness campaigns did not: it has given emissions numbers an owner, a deadline, a checker and, under the CCTS, a price. The resulting demand is concentrated, operational and sector-specific. It sits in a few hundred large plants in eight named sectors, in the finance and secretarial functions of the top 1,000 listed companies, and in the export desks of producers selling covered goods to Europe. Ventures that start from plant data and work outwards to disclosure and border compliance are better placed than those that start from reports and try to work inwards. The rules are still moving, particularly on value chain disclosure and offset methodologies, so the map should be read as a snapshot of October 2026, to be revisited as each notification lands.
Methodology and sources
This article is based on published rules and official or reputable summaries, opened in October 2026. The BRSR Core circular details come from SEBI's circular page (https://www.sebi.gov.in/legal/circulars/jul-2023/brsr-core-framework-for-assurance-and-esg-disclosures-for-value- chain_73854.html) and the circular PDF (https://www.sebi.gov.in/sebi_data/attachdocs/jul-2023/1689166456465.pdf), with the list of nine attributes and their KPIs from KPMG India's First Notes of 15 July 2023 (https://assets.kpmg.com/content/dam/kpmg/in/pdf/2023/07/firstnotes-sebi-framework-on-brsr-core-value-chain-disclosures-assurance-by- listed-entities.pdf). The December 2024 easing of value chain disclosures is taken from the editor-verified source list and should be checked against SEBI's current position. The CCTS notification date and mechanism structure come from the IEA policy database (https://www.iea.org/policies/25639-carbon-credit-trading-scheme-2023) and the Bureau of Energy Efficiency's carbon market page (https://beeindia.gov.in/en/programmes/carbon-market). The October 2025 targets and sector reductions come from ETV Bharat (https://www.etvbharat.com/en/!bharat/india-notifies-first-emission-intensity-targets-for-carbon-intensive-sectors-enn25101002607); the January 2026 expansion from The Statesman (https://www.thestatesman.com/business/govt-notifies-greenhouse-gas-emission-intensity- targets-for-208-more-carbon-intensive-industries-1503544850.html); the smelter example and penalty structure from JSA's April 2025 note on the draft rules (https://www.jsalaw.com/wp-content/uploads/2025/04/JSA-Prism-Climate-Change-GHG-EIT-Rules.April-2025.Final_.pdf). CBAM scope and dates are from the European Commission (https://taxation-customs.ec.europa.eu/carbon-border-adjustment- mechanism_en). National emissions figures are from Down To Earth's summary of India's Fourth Biennial Update Report (https://www.downtoearth.org.in/amp/story/climate-change/energy-and-agriculture-sectors-lead-emissions-in-indias-latest-greenhouse-gas- inventory). The demand map, the pull-strength classification and the ordering of product categories are the author's analysis, not observed market data. PanScience and PredCo descriptions are company-reported.