India coal outlook: from tonnes to delivered fuel security

The commercial question has changed

An Indian coal buyer has at least three different problems: obtaining enough physical fuel, obtaining fuel that performs in the boiler or process, and obtaining it at a cost compatible with the business’s output price. A national production milestone can improve the first condition without resolving the other two. A procurement strategy built around one aggregate demand forecast therefore misses much of the commercial risk.

The working thesis of this report is that the next increment of value comes from reconciling the supply chain rather than collecting more unqualified headlines. The buyer should connect mine-level commitments to dispatch capability, the nominated route to actual transit variability, and the invoice specification to the quality received at the plant. Each reconciliation changes the interpretation of a nominal discount.

The official fiscal-year record illustrates why segmentation matters. Domestic coal output was 1040.08 MT in FY 2025-26 against 1047.52 MT a year earlier. Separately, captive and commercial blocks produced 210.46 MT. These are observations about production and ownership composition, not delivered availability for a particular consumer.[S2] A commercial mine selling to third parties and a captive block serving its owner have different implications for an external buyer’s addressable supply.

Read the annual balance without mixing scopes

The important analytical result is the divergence, not a claim that one number is more correct. Overall imports can rise while a large end-use subset falls because other categories or users move differently. The annual category table and the thermal-plant disclosure have separate boundaries. Neither permits an exact national consumption balance without exports, stock movements, coverage reconciliation and calorific-value adjustments.

The Ministry’s production webpage also retains a historical April-November production panel next to a more recent annual import series.[S1] A dashboard that simply scrapes every visible table as the same observation date can generate a misleading comparison. The minimum data model needs a period start, period end, unit, commodity definition, reporting boundary and revision flag for every series.

For commercial monitoring, separate thermal power, coking coal, industrial non-coking demand and inventories. An annual fiscal series is useful for structural exposure; it is too slow for a cargo nomination or an intramonth fuel-security decision. The appropriate next observation is usually a plant, route or shipment event rather than another national total. This distinction also prevents an unjustified extrapolation of one fiscal-year change into an assured future trend.

A reproducible delivered-heat model

Comparing invoice prices per tonne silently assumes that each tonne contains the same usable energy and produces the same operating result. It does not. The official grading system uses gross calorific value for non-coking coal and different quality measures for coking coal.[S3] Thermal comparisons should begin on a common, contractually defined measurement basis. An as-received number cannot be substituted uncritically for an air-dried certificate.

The resulting costs areINR 2.86/kWh for the domestic case andINR 3.62/kWh for the imported case, rounded. At the same assumed efficiency, the domestic option is cheaper on energy, even though a higher-quality imported tonne contains more heat. Solving the equation in reverse gives an imported CFR break-even ofUSD 67.56/t at the stated exchange rate and logistics cost. This threshold is conditional, not a price forecast.

Before converting this screen into a purchase decision, replace every assumed value with contemporaneous bids and independently reconcile moisture, ash, handling constraints, efficiency, credit terms and rejection rights. Higher heat content may reduce tonne movements, but operational incompatibility or a quality dispute can dominate that saving. A robust model records these adjustments as explicit items rather than hiding them in a miscellaneous premium.

The scarce product may be a reliable route

A coal purchase has a geographical identity. Mine availability, siding readiness, rake placement, corridor congestion, unloading and last-mile handling are linked operations. An attractive ex-mine quotation is only useful if the buyer can move the contracted quantity on the required days. Conversely, an alternative route can be valuable even when its average freight is higher, if it reduces a material interruption exposure.

The July 2026 government update reported 72 commissioned first-mile-connectivity projects with about 589 MTPA handling capacity, within a larger 139-project programme.[S5] Those capacities are infrastructure indicators. They should not be treated as incremental coal production or automatically accessible capacity for every consumer. A project can remove one bottleneck while a downstream junction remains constrained.

Enalyt’s suggested route register contains the source mine, loading point, rail corridor or modal chain, contractual handover point, dispatch lead time, median and adverse transit time, unloading capacity and inventory buffer. The adverse transit measure should be defined from observed delivery distributions. Where no sufficient history exists, use a clearly labelled stress assumption and revise it when evidence improves.

Route diversification must also pass a correlation test. Two suppliers using the same constrained siding may provide less resilience than their supplier count suggests. A coastal alternative adds port and transshipment interfaces; it should be assessed on the full chain, including cargo size, berth access, inland connectivity and demurrage allocation. This is a portfolio problem: minimise common failure points rather than multiplying vendor names.

Inventory: reconcile actual days with normative days

Stock days are a ratio, so changing the denominator can change the headline without moving a single tonne. The historical CEA framework published by the Power Ministry sets month-varying norms using 85%PLF, with different ranges for pithead and non-pithead plants.[S6] Operational dashboards also calculate coverage against actual or planned burn. A well-run dashboard shows both and identifies the applicable plant category and month.

The carrying-cost calculation supplies one side of the decision. The other is the expected cost of fuel-related interruption, including lost contribution, replacement purchases, contractual consequences and recovery time. Those values are plant-specific; this report does not invent them. A procurement committee should compare inventory cost with a documented interruption scenario and identify who accepts the residual exposure.

A snapshot also needs an observation date. The Power Ministry disclosed 42.8 MT of thermal-plant stock on 12 July 2026, described as 14 days at 85%PLF.[S8] That is a dated national observation, not a live November stock assessment. The useful operating signal is the trajectory of usable stock, confirmed incoming deliveries and expected burn over the next planning window.

Coking coal is a different exposure

The same fuel-market narrative should not be applied mechanically to steel. A thermal buyer screens energy and boiler performance; a coke-making or blast-furnace supply chain must also preserve metallurgical blend performance. Domestic raw output, washery input, clean-coal yield and qualified blend tonnes are separate stages. Adding raw tonnes does not establish an equivalent reduction in imported blend requirements.

The January 2026 government notification designated coking coal as a critical and strategic mineral.[S7] The investment implication is a need to examine the conversion chain: geological quality, mining access, washing performance, logistics and customer qualification. The designation itself is not evidence that a mine has been commissioned, that washing economics are attractive, or that a steel producer will accept the product.

A practical procurement model should therefore use a qualified-product bridge. Start with raw tonnes, apply saleable yield, then qualification acceptance, then the share that can replace a specific imported grade without degrading the process. Keep each rate visible. A lower raw-coal price can be offset by low recovery, disposal costs, lost yield or a narrow usable blend envelope.

This also changes scenario design. An import-exposure reduction might come from more washed domestic material, a change in steel production route, improved process efficiency, or a weaker output cycle. Each has a different competitive consequence. A scenario that assumes a national output target automatically displaces imported hard coking coal is omitting the very constraints that determine commercial value.

Outlook: three conditional pathways, not one point forecast

The following transparent sensitivity shows how different annual growth assumptions change a 2026 index through 2030. The exercise is designed for procurement planning: it makes the growth assumption visible and allows the reader to replace it. It is not a probability-weighted forecast or an official outlook.

In a low-growth case, coal tonnage may be relatively flat while peak-hour system dependence remains material. Lower annual volume does not automatically remove the need for prompt dispatch, especially if plant operations become more variable. In a higher-growth case, mine capacity, quality and logistics all need to be tested; one aggregate supply-growth assumption is insufficient.

The procurement response should depend on the exposure that changes. A short-lead domestic buyer may prefer contractual flexibility if its production cycle is volatile. A long-route consumer may value confirmed transport and buffer stock more than a marginal price discount. A company exposed to imported metallurgical coal should maintain a separate currency and grade-risk dashboard even if thermal import dependence is falling.

The signals that would change this assessment are identifiable: sustained changes in plant burn, reliable route-level delivery improvements, the commissioning and actual throughput of processing infrastructure, and qualified substitution at industrial consumers. Monitoring these signals is more defensible than repeatedly revising a headline price target without a disclosed model.

A 90-day agenda for procurement and strategy teams

The first month should establish a common data dictionary. Reconcile fiscal years, coal versus lignite, production versus dispatch, mine ownership categories, customs categories, plant receipts and stocks. Record source publication dates separately from observation dates. Assign an owner to each recurring series and require explicit sign-off when a definition changes. This work is unglamorous, but it prevents false alerts and unearned conclusions.

The second month should connect contract economics to operations. Build a delivered-heat comparison for each material supply lane, using actual quality and current commercial offers. Reconcile cash taxes, recoverable taxes, credit terms and transportation responsibilities with the finance team. Add a stock-days bridge on both actual and normative bases and show how different operating schedules change cover.

The third month should stress the portfolio. Test a delayed rake, a quality downgrade, a currency movement, an imported cargo disruption and a change in process demand. Identify which exposures occur together. Document the procurement alternatives that remain operationally available and the lead time needed to activate them. Resilience only has value when the alternative can actually be executed.

• For procurement heads: approve source-route-quality alternatives, not just a supplier shortlist.
• For operations: reconcile usable inventory with the production schedule and quality restrictions.
• For finance: track delivered-cost sensitivities, working capital and foreign-currency exposure separately.
• For strategy: distinguish structural substitution from temporary stock or dispatch effects.

The resulting database should produce an audit trail from each management conclusion back to a dated observation or a clearly stated assumption. That is the foundation for recurring market intelligence: a repeatable decision process rather than a larger collection of disconnected charts.