Heat pump versus boiler: how to work out whether the switch pays on your plant
Whether an industrial heat pump beats your boiler is decided by one number, the breakeven COP: your electricity price per kilowatt-hour divided by what a useful kilowatt-hour of heat costs you from the fuel you burn today. In India in 2026 that number is 1.0 to 1.2 against piped natural gas and furnace oil, which means any working heat pump wins, and 3.3 to 6.5 against coal, which means a heat pump on grid power generally does not. Tetra Heat Pump publishes the calculation rather than a savings percentage, because the percentage depends entirely on the fuel being replaced and on what the plant pays for power.
What number actually decides this?
The breakeven COP is the efficiency at which a heat pump exactly matches what your boiler costs to run. Below it you lose money; above it, every point of COP is margin. You hold both inputs already:
Breakeven COP = your landed electricity price per kWh ÷ the delivered cost of a useful kWh of heat from your current fuel.
No vendor model and no hidden assumptions. Everything below exists to help you get those two inputs right.
| Fuel burned today | Breakeven COP in India, 2026 | What it means | The argument that actually applies |
|---|---|---|---|
| Piped natural gas | 1.0 to 1.2 | Any working heat pump beats it on fuel cost | The fuel bill |
| Furnace oil | 1.0 to 1.2 | Any working heat pump beats it on fuel cost | The fuel bill |
| Coal | 3.3 to 6.5 | A heat pump on grid power does not beat it on fuel cost | Compliance, air quality, labour, customer audit and power procurement |
Light diesel oil, diesel and LPG sit on the same side of the line as gas and furnace oil. Biomass briquette sits with coal.
Segment by the fuel being replaced, not by the industry you are in. Two food plants in one cluster, one on PNG and one on coal, are not the same prospect. Your industry decides the temperatures you need; your fuel decides whether the economics work at all. If you are on gas, there is more market context on the Indian industrial gas price and a version of this comparison run against a PNG boiler alone; if you are on oil, the furnace oil version is here.
How do I work out what a kilowatt-hour of heat costs me today?
Three inputs, then one division. Take the delivered price of the fuel in the unit you are invoiced in, the calorific value of the grade you are supplied, off the supplier’s certificate and not a textbook, and your boiler’s efficiency on the same basis. Divide the price by the product of the other two, then convert to kilowatt-hours. This is where the arithmetic usually goes wrong.
Gross and net calorific value are not interchangeable. Suppliers quote one, boiler manufacturers often quote efficiency against the other, and an efficiency taken on net looks better than the same boiler measured on gross. Mix the bases and the answer comes out wrong in whichever direction flatters the technology being sold. Pick one basis and use it on both sides.
The nameplate efficiency is not your efficiency. The plate figure is a combustion efficiency measured on a test day. What you want is annual fuel purchased divided by useful heat delivered to the process, because in an older plant the two are separated by losses nobody puts in a comparison table:
- Standing loss. A boiler held hot overnight and over the weekend so it is available on Monday burns fuel to heat nothing.
- Blowdown. Energy leaves the plant in hot water at boiler pressure, with the treatment chemicals it was dosed with.
- Distribution and condensate return. Steam mains, un-lagged flanges, failed traps and condensate going to drain instead of back to the hotwell all cost fuel, and none of it shows on the nameplate.
- The temperature you paid for and threw away. A boiler raising saturated steam at pressure to serve an 80°C wash tank through a pressure reducing station made high grade heat, paid for it in fuel, then degraded most of it across a valve. That is the load a heat pump is best at. It is also the load least visible in a plant’s own energy accounting.
What Indian industrial fuels cost
| Fuel | Price | As of | Note |
|---|---|---|---|
| Light diesel oil | ₹107.65 per litre | 1 August 2026 | |
| Furnace oil | ₹71.99 per kg, ₹69.07 per litre | 1 August 2026 | Visakhapatnam depot reference |
| PNG, industrial, Gujarat general | ₹68 per SCM | June 2026 | An industrial contract rate. Published domestic piped gas rates are a different and much lower product, and must not be used for factory arithmetic |
| PNG, industrial, Morbi ceramic cluster | ₹75 per SCM | June 2026 | A cluster premium on the same fuel |
| Biomass briquette, 60 mm | ₹8,000 to ₹10,000 per tonne | 2026 | Trade pricing, not an official index |
| Coal, non-coking G13 | ₹1,849 per tonne | June 2026 | Pithead notified price only. Delivered cost at a factory gate is materially higher, and freight, handling and ash disposal all sit on top |
That is input one. The other two come off your own paperwork, because nobody else can supply the calorific value of the grade delivered to your gate or the efficiency of your boiler. It is also why there is no filled-in example here: any example would have to invent both, and an invented example is exactly what you should be checking a vendor for. There is no LPG or propane row either, because no defensible 2026 Indian industrial figure was available. Use your own delivered price.
And what should I put in for electricity?
The landed rate, not the energy charge on the tariff order. Demand charges are billed separately per kVA per month and are part of what a delivered kilowatt-hour costs you, as are the fuel and power purchase adjustment, electricity duty and any cross-subsidy surcharge. Dividing by the headline energy charge overstates the heat pump’s case.
| State and licensee | Category | Energy charge | Demand charge | Landed |
|---|---|---|---|---|
| Maharashtra, MSEDCL | HT-I | ₹8.44 per kVAh | ₹650 per kVA per month | Add duty and any time-of-day adjustment |
| Gujarat, GUVNL | HT-I | ₹4.30 per unit | Not separately verified | ₹7.50 to ₹8.50 all-in |
| Tamil Nadu, TANGEDCO | HT industrial | ₹7.50 per kWh | ₹608 per kVA per month | Plus 5 percent electricity tax |
| Karnataka, BESCOM and the other ESCOMs | HT-2(a) | ₹6.70 per unit plus 35 paise surcharge | ₹365 per kVA per month | About ₹7.05 before duty |
| Uttar Pradesh, UPPCL | LMV-6, Small and Medium Power | ₹9.02 per unit average billing rate | Not verified | Already all-in |
The Uttar Pradesh figure is LMV-6 Small and Medium Power, not a large HT industrial tariff, so do not read it as the rate a big plant would pay. Maharashtra bills HT-I per kVAh, not per kWh, so a plant running poor power factor pays more per useful kilowatt-hour than ₹8.44 suggests, and any Maharashtra calculation has to say so.
The Maharashtra row also holds the only forward price signal in this comparison. The approved HT-I path declines from ₹8.44 per kVAh in FY2026-27 to ₹7.45 by FY2029-30. One side of your comparison sits on a published, regulated, downward track; the other tracks a commodity you do not control.
What COP will I actually get, and at what temperature?
COP is set by the lift, the gap between source and output temperature, not by the output alone. A COP quoted without both conditions cannot be used in a comparison, and an engineer is right to discount one that arrives bare.
Independently measured performance in Indian conditions at the hard end of the range is about 2.26 at 110°C output and about 2.0 at 120°C. Set those against the breakeven table. Even at the top of the envelope, where the machine works hardest, both sit well clear of the 1.0 to 1.2 needed to beat gas or furnace oil and well short of the 3.3 to 6.5 needed to beat coal. That is the whole economic argument, and it needs no savings percentage.
Performance improves as the required output falls. Tetra Heat Pump water source units reach COP 4.5 to 5.2 at 60°C output, and performance falls as output temperature rises toward the 120°C ceiling. Both ends of that curve are real, and quoting only one is how the category lost credibility with plant engineers. The money in most Indian plants sits in the lower bands anyway, because that is where the hot water volume is.
The source is the other half of the lift, and the input buyers most often leave on the table. Warm effluent, condenser water, compressor jacket heat or a refrigeration plant rejecting heat while the process needs it all beat ambient air, and a better source means a shorter lift and a higher COP. Our notes on how a heat pump works and on air source versus water source cover both.
Tetra Heat Pump, a product of Promethean Energy Private Limited, builds air source, water source and cascade units delivering continuous output up to 120°C with ±0.5°C control, which covers hot water loops, wash and pretreatment baths, clean-in-place circuits and aqueous jacketed vessels, and none of the duties above that ceiling.
Then why does everyone quote a savings percentage?
The reason is that a percentage is far easier to print than a method, and the figure quoted is almost always the best case.
A savings percentage is a function of the fuel displaced, the output temperature and the electricity tariff. Change one of the three and the same machine on the same site gives a different answer. A page carrying a blanket percentage is therefore either quoting the most favourable combination it could find or not saying which fuel it displaced, and neither survives the obvious follow-up. The breakeven COP travels better, because it is a ratio between two prices you can verify.
What is the payback, and how should I test the one I am quoted?
Two lines of arithmetic:
Annual saving = annual thermal demand in kWh × (cost per useful kWh today, less cost per useful kWh with the heat pump).
Simple payback = installed capital cost ÷ annual saving.
Run it on your own numbers, then reverse it on the proposal in front of you. If a quoted payback does not reconstruct from your thermal demand, your tariff, your fuel price and a COP stated with its source and output conditions, the difference is somebody’s assumption and you are entitled to see it.
Then there is the finding that reframes the conversation. Indian industrial buyers consistently say they want 1 to 3 years. The realised average across 224 Indian MSMEs in the World Bank study was 44 months. Both are true at once, because the stated preference is a screening filter used to reject proposals early rather than a hurdle rate. A proposal that arrives with the working shown and sourced survives a longer number than one that arrives with a round percentage.
Capital allowances change the cash profile, not the saving. Energy saving devices carry 40 percent accelerated depreciation and heat pumps are named explicitly in the income tax rules, as our guide to Indian incentives sets out.
What about the capital cost difference?
A heat pump does cost more to install than an equivalent fired boiler, and any comparison that conceals this fact is not worth reading.
The fair version is fully installed and commissioned on both sides. On the boiler side that means the items buyers treat as sunk because the last boiler came with them: fuel storage and handling, the chimney and flue gas path, statutory approvals and registration, feedwater treatment, and the floor space all of it occupies. On the heat pump side it means the electrical work, the source side pipework and the tie-in. Compare project to project, not machine to machine.
We publish no price bands, because a band wide enough to be honest is too wide to be useful. What drives the number is the duty, the output temperature, the source available, the electrical work and how much existing plant is reused. Apply the arithmetic above to the quotation you are holding, and use the calculator to test it at different output temperatures.
Where does a boiler still win?
A boiler still wins in four situations, and none of them is an edge case.
Above the heat pump’s temperature envelope. Curing and drying ovens, calcination and reheat furnaces are combustion duties and will stay combustion duties.
Saturated steam at pressure. If the process needs steam because it needs steam, a heat pump is not the answer, whatever the hot water loop beside it is doing.
Very intermittent duty. A machine sized for a short sharp peak runs badly at part load the rest of the day. That is a design objection, not a price one, and it kills more projects than cost does.
Fuel cost against coal. The 3.3 to 6.5 breakeven band is the reason, and it is not close. Levelised, a heat pump reaches roughly parity with coal on grid power and is clearly ahead only where the site has captive renewable generation. Claim fuel savings against coal on a standard industrial tariff and the first competent energy manager to read it will say so.
| Decision factor | Heat pump | Fired boiler |
|---|---|---|
| Output temperature | Strong in the low and mid bands, and less efficient as the required output rises | Indifferent to output temperature, and the only option above the heat pump’s ceiling |
| Load profile | Best on steady, predictable base load | Absorbs sharp intermittent peaks without an efficiency penalty |
| Fuel displaced | Wins on running cost against gas, oil and diesel | Wins on running cost where the fuel is coal or briquette |
| Capital available | Higher installed cost | Lower installed cost |
| Statutory burden | A hot water loop below 100°C is not a registered boiler under the Indian Boiler Regulations | Registration, a certified attendant across shifts and an annual inspection shutdown where the regulations apply |
| Backup and resilience | Depends on the electrical supply and on sanctioned load headroom | Depends on fuel delivery, storage and the price you are exposed to |
Should I replace the boiler or run both?
In most Indian retrofits you should run both. This is usually the better project rather than the more cautious one.
The shape that works is a heat pump taking the base load below 100°C while the fired system stays for the top end and for backup through at least the first year. It wins on risk, because the plant is never one machine away from a production stop; on payback, because the capital goes to the hours the machine runs rather than to a peak it meets a few times a month; and on measurement, because you can compare the heat pump against the boiler it displaces on the same site, in the same season, with both available.
Sizing follows from the load profile, not the peak. Plot hourly thermal demand for a representative month and a machine covering well under the peak usually still covers most of the annual kilowatt-hours. Sizing to the peak buys capacity that spends its life at part load, which is where COP degrades and the project quietly stops paying.
The practical constraint is the tie-in, because most plants run one header serving loads at very different temperatures and the low temperature duties have to be separated out first, which is the whole of the question on a thermic fluid heater. That is a piping and controls exercise before it is a purchase. There is more on the equipment and on integrating with existing plant, BMS and controls.
What changes in operating cost besides the fuel?
Maintenance differs in kind more than in amount. A boiler needs tube cleaning, burner service, water treatment, flue gas testing and combustion tuning; a heat pump needs the refrigerant circuit checked, the heat exchangers kept clean and the controls calibrated. A vendor telling you it is maintenance free is describing a machine that gets measured for COP once and never again.
The recurring statutory cost is the real difference, and most comparisons omit it. A hot water loop below 100°C is not a registered boiler under the Indian Boiler Regulations, so the certified attendant on every shift and the annual inspection shutdown go with it, and both are permanent line items in a plant budget. The exemption applies below 100°C only, so scope it against the duty you are moving, not the whole boiler house.
There is also no fuel to store, handle, weigh or lose to pilferage, and no ash to dispose of. And the operating cost stops tracking a commodity, because you exchange a fuel price set in a global market for one set in a published state tariff order.
What does switching do to my electricity connection?
It raises your sanctioned load, and this is the part no vendor page in India tells you. Moving a thermal load off fuel and onto electricity increases the plant’s connected and contracted demand, which can trigger a connection or load enhancement charge and, more painfully, a revision of your security deposit.
In Maharashtra, security is set at twice the average billing cycle and recalculated annually, so a heat pump that raises the monthly electricity bill by ₹6 lakh pulls through roughly ₹12 lakh of additional deposit at the next recalculation. That is working capital rather than project cost, and it lands after commissioning. In Gujarat, where additional demand is sanctioned, the deposit is computed as though it were a new service for the entire load rather than for the increment, so the exposure is larger again.
Check twelve months of your own billing for headroom before assuming the worst. Many plants carry sanctioned load they no longer use, left over from equipment retired years ago, and the increment fits inside it. Where it does not, your licensee publishes a schedule of connection and processing charges worth pulling before the project is budgeted.
Running the comparison on your own plant
Take your landed electricity rate per kilowatt-hour from your bill, and your delivered cost of a useful kilowatt-hour of heat from the table above plus your own calorific value and boiler efficiency. Divide the first by the second. That is your breakeven COP, and it settles the question before any equipment is specified.
If it sits below the COP achievable at the duty you want to move, the next step is a look at your load profile and source options, not a quotation. You can get that started here.
Frequently asked questions
- How do operating costs compare between a heat pump and a boiler?
- Operating cost is decided by the breakeven COP: your electricity price per kilowatt-hour divided by what a useful kilowatt-hour of heat costs from your current fuel. In India in 2026 that ratio is 1.0 to 1.2 against piped natural gas and furnace oil, so a heat pump operating above the ratio costs less per unit of heat delivered than the boiler it replaces. Against coal the same ratio is 3.3 to 6.5, which a heat pump on grid power generally does not reach.
- What COP do I need to beat my current fuel?
- Between 1.0 and 1.2 if you burn piped natural gas or furnace oil, which almost any working industrial heat pump clears. Between 3.3 and 6.5 if you burn coal, which a high temperature machine on grid power generally does not. Work out your own figure by dividing your landed electricity tariff by the delivered cost of a useful kilowatt-hour of heat from the fuel you actually buy.
- Does a heat pump beat a coal boiler?
- It does not pay on fuel cost when run on grid power. The breakeven COP against coal is 3.3 to 6.5, which sits above what a high temperature industrial heat pump achieves, and levelised the two come out at roughly parity on grid power with the heat pump clearly ahead only where the plant has captive renewable generation. Against coal the case worth examining is compliance, air quality, labour and customer audit, not the fuel bill.
- Can a heat pump replace a steam boiler?
- It cannot replace a boiler raising saturated steam at pressure. What it does replace are the low temperature duties that a steam boiler is currently serving through a pressure reducing station, which in many plants is a large share of the fuel burned, and the steam boiler stays for the loads that genuinely need steam. A hot water loop below 100°C also falls outside the Indian Boiler Regulations, which removes the certified attendant across shifts and the annual inspection shutdown for that part of the plant.
- What payback should I expect?
- Calculate it rather than accept one. Annual saving is your annual thermal demand multiplied by the difference in cost per useful kilowatt-hour, and payback is the installed capital cost divided by that saving. Indian industrial buyers say they want 1 to 3 years, but the realised average across 224 Indian MSMEs in the World Bank study was 44 months, so treat the shorter figure as a screening filter rather than a hurdle rate.