India gas intelligence: the gap between infrastructure and economic demand

The growth thesis needs an affordability test

The first distinction for an Indian gas strategy is physical access versus economic demand. A pipeline reaching a geographical area creates the option to consume gas; it does not guarantee that every customer can switch profitably, that all connections are active, or that gas is available on the required commercial terms. An LNG terminal can add regasification capability while local demand remains price-sensitive. A useful market view therefore links infrastructure to customer economics and contract flexibility, rather than treating kilometres and nameplate tonnes as a direct demand forecast.

PNGRB’s FY2025-26 statistical report shows gas consumption of 68.542 bcm and LNG imports of 34.216 bcm in its aggregate table, compared with consumption of 71.314 bcm in FY2024-25. That table’s definitions and reported production balance should be retained with the observations. The IEA’s 2025 report projects 103 bcm of calendar-year consumption in 2030 under its forecast assumptions. These are different periods and statistical contexts: a historical contraction does not disprove every long-term growth scenario, and a long-term projection does not turn the following year into a guaranteed recovery. [S1] [S2]

Build one reconciled gas balance before modelling the future

Gas datasets are unusually vulnerable to false precision. Gross field production, net production after flaring and losses, LNG imports, sales and consumption are not always reported on one identical boundary. Energy and volumetric units introduce another layer: a tonne of LNG, a standard cubic metre of gas and a million British thermal units measure different things. Conversion requires a stated calorific value or composition assumption. The ministry’s statistical compendium and PNGRB’s regulatory report should therefore remain separate source vintages unless their definitions are explicitly reconciled. [S1] [S10]

For a commercial database, attach the gas basis, temperature and pressure reference where available. Preserve whether a price is on gross or net calorific value, and whether a volume is measured before or after regasification or distribution losses. A small conversion inconsistency can matter when bids compete on a narrow delivered-price difference. Data quality is not an appendix to the investment case: it determines whether the fuel-cost comparison is actually comparing the same useful energy.

The operational balance should also distinguish contracted availability from physical flow. An annual LNG supply agreement is not the same as a confirmed cargo delivery date, and a regasification reservation is not evidence that downstream transmission capacity is available at the same time. Track the chain from source commitment through berth, storage, send-out, pipeline and customer nomination. That exposes the bottleneck that can prevent a physically abundant molecule from reaching a commercially attractive customer.

A demand-to-import bridge with explicit assumptions

The scenario table begins with PNGRB’s 68.542 bcm FY2025-26 consumption observation and compounds assumed growth for four fiscal intervals to FY2029-30. It then subtracts an independently assumed domestic supply contribution. These inputs are not fitted forecasts and should not be confused with the IEA’s calendar-2030 projection. Their purpose is to reveal what an LNG procurement requirement would look like under different combinations of demand and domestic availability. The residual is a simplified import requirement, not a cargo schedule or a precise trade forecast.

The important sensitivity is not just the demand endpoint but the distribution of that demand. An incremental industrial customer with an alternative fuel can reduce gas offtake when the spread changes. A customer using gas as a process feedstock may have a different response. Procurement commitments should be matched to the resilience of the underlying load, not only its expected annual volume. A growth case supported by highly interruptible demand warrants different contract flexibility from one supported by stable, creditworthy baseload.

The delivered LNG stack: avoid the FOB versus DES trap

The starting point in our cost model is an assumed delivered-ex-ship LNG commodity price of $10/MMBtu. Because that price already includes the voyage to the receiving terminal under the assumed contract basis, we do not add another full ocean-freight charge. We then assume regasification of $0.60, transmission of $0.65 and marketing/service costs of $0.20 per MMBtu. The resulting gas-gate cost is $11.45/MMBtu before taxes and any local distribution or customer-specific charges not included in those inputs. None of these amounts is a current tariff or executable quote.

A FOB offer requires a different bridge, including shipping, voyage fuel and any relevant loss allowance that is not already priced. A seller’s commodity formula may reference oil, a gas hub or a negotiated fixed price; a buyer’s effective cost can also include take-or-pay exposure and unused capacity. Domestic gas is not interchangeable with an LNG spot quote. The April 2023 pricing notification established a specific APM framework, while PPAC publishes applicable price notices; eligibility, allocation, category and current amendments need verification for the actual buyer. [S6] [S7]

This means a quoted gas price needs metadata. Record the delivery point, energy basis, index, averaging window, currency, tax treatment and contractual quantity obligations. A procurement dashboard that ranks an upstream regulated price against a fully delivered industrial tariff without those fields will identify false bargains. The correct comparison is customer-gate cost for a common date, volume profile and useful-energy requirement.

Industrial fuel switching: compare useful heat, not nominal units

Consider an entirely hypothetical fuel-oil alternative priced at INR 45/kg, with an assumed calorific value of 40 MJ/kg and a boiler efficiency of 82%. Its useful-heat cost is approximately INR 1,447.49 per MMBtu. At 90% gas efficiency, the base delivered-gas assumption of INR 973.25/MMBtu becomes INR 1,081.39 per useful MMBtu. The apparent gas advantage is approximately INR 366.10 per useful MMBtu before burner-conversion investment, incremental distribution charges, taxes, downtime and operating differences. The calculation is useful precisely because every input can be challenged.

Under those assumptions, delivered gas reaches parity at about $15.33/MMBtu. Subtracting the assumed non-commodity gas costs gives a DES commodity threshold near $13.88/MMBtu. This is a sensitivity result, not a universal switching price. It changes with foreign exchange, alternative-fuel quality, actual efficiency and the cost of a disrupted process. The procurement decision should also consider whether the plant can switch back, whether operators can maintain quality, and whether the gas contract permits lower offtake when the fuel spread reverses.

Power generation needs its own conversion. At an assumed heat rate of 7.5 MMBtu/MWh, $11.45 gas and INR 85/$, the fuel-only cost is about INR 7.30/kWh. That excludes fixed charges, start-up fuel, part-load efficiency, transmission and other costs. It is not a forecast of an electricity exchange price or a claim about an existing plant. It demonstrates why a national gas-growth thesis should not assume every unit of gas-fired capacity will dispatch economically at every LNG price.

Connections and terminal capacity need denominator discipline

PNGRB’s 31 March 2026 CGD snapshot reports 1.69 crore domestic PNG connections but 1.07 crore billed connections. Those counts imply a billed-to-connected ratio near 63% using rounded inputs. They do not establish that every unbilled connection is physically disconnected or has zero potential demand: billing periods, reporting conventions and customer activation matter. The October 2025 MIS also separates targets, actual connections and billed connections, reinforcing the need to retain each field rather than use a single network headline. [S1] [S5]

Terminals have a related denominator problem. Petronet’s Q1 FY2026-27 release explicitly states that Dahej’s nameplate base increased from 17.5 to 22.5 MMTPA. Its reported utilisation therefore cannot be interpreted against an earlier percentage without adjustment. The commissioning notice dates the additional facilities to 31 March 2026. Compare physical TBTU, the available capacity denominator, period length and contracted availability together. An expanded denominator can lower a utilisation percentage even when the change in physical throughput is much smaller. [S3] [S4]

The practical implication is to model a network as a chain of constraints. Nameplate capacity is only the first screen. A buyer also needs reliable berth access, storage, send-out capability, a matching pipeline route, distribution capability and customer offtake. Each constraint can have a different counterparty and charging structure. A national capacity sum cannot establish deliverability to one plant, just as a map of authorised geographical areas cannot establish that a customer’s connection is commissioned and economic.

Commitment risk: when growth assumptions become payment exposure

A simplified take-or-pay stress makes the issue concrete. Assume a contractual daily quantity of 100,000 MMBtu and an 85% minimum obligation, with actual demand falling to 60,000 MMBtu/day for 30 days. The arithmetic shortfall against the assumed floor is 25,000 MMBtu/day, or 750,000 MMBtu for the month. At an assumed $10/MMBtu commodity price, gross exposure is $7.5 million before make-up rights, resale, mitigation, scheduling rules and the actual settlement formula. That amount is not automatically a payable invoice under every contract.

The useful procurement question is how much demand can fall before the business crosses its quantity floor, and how quickly it can respond. Keep customer-load sensitivity, supplier commitment, terminal reservation and pipeline booking in the same stress model. If contracts settle on different timing or energy bases, a perfectly balanced annual portfolio can still create monthly cash pressure. A cheaper long-term slope should be evaluated alongside flexibility, destination options, credit and the resilience of the underlying end-use load.

A practical research agenda for gas buyers and developers

Start with customer economics rather than a market-size target. Segment demand into process-critical, economically switchable and discretionary loads, then test each against delivered gas prices and operational constraints. A supplier can use this to design flexibility; a network developer can use it to prioritise activation rather than simply connections; an industrial buyer can use it to negotiate the right balance of cost and optionality. The evidence required is granular but not mysterious: comparable energy prices, measured efficiency, actual load shape, current capacity availability and contract terms.

For market surveillance, maintain separate observations for national supply, LNG imports, terminal throughput, billed connections, gas price categories and customer-sector activity. Date-stamp each dataset and preserve revisions. Research should explain a change before extrapolating it: lower imports may reflect demand, domestic availability, inventory, price response or timing. The source period and boundary should be visible wherever a number is used. That habit is the difference between a useful intelligence service and a plausible-looking collection of growth claims.

The forward view remains constructive only under conditions that can be tested. Infrastructure must become usable capacity, customer-gate pricing must support competitive operations, and procurement commitments must remain robust when demand disappoints. The scenarios here identify those dependencies rather than promise a market trajectory. A company can replace the assumptions with its own operating evidence and reach a different conclusion while retaining the same transparent analytical method. That is the purpose of the trial report.