China's 15th Five-Year Plan: What Year One Shows

Increments on carbon, silence on dates, and one deletion that matters

The 15th Five-Year Plan was published on 13 March 2026 and covers 2026 to 2030. Its energy and climate content is best read as three layers. The binding numbers are a 17 per cent reduction in carbon intensity across the period and a non-fossil share of primary energy consumption of 25 per cent by 2030, up from 21.7 per cent in 2025. The capacity ambitions are 100 GW of offshore wind and 110 GW of nuclear by 2030. The qualitative commitments are to promote the peaking of coal and oil consumption, to continue the clean and efficient utilisation of fossil fuels, and to expand coverage of the mandatory carbon market.

The deletion is more informative than any of the additions. Previous plans carried a binding energy intensity target, a reduction in energy consumed per unit of GDP. That target has been removed from the binding set. Carbon intensity and energy intensity constrain different things. Carbon intensity can be satisfied by changing the fuel mix while total energy consumption grows. Energy intensity constrained the growth of consumption itself. Removing it loosens the ceiling on absolute energy demand at exactly the moment electrification and data centre load are pushing that demand up.

Li Shuo of the Asia Society Policy Institute characterised the carbon intensity target as a quiet recalibration, indicating how difficult the original 2030 goal has become. That reading fits the structure of the document. This is not a plan written by a government that believes its previous trajectory was achievable on schedule, nor one that intends to abandon the direction. It is a plan that buys headroom.

For the hydrocarbon chapter the language is unambiguous and runs the other way. Annual crude production is to be held at around 200 million tonnes. Gas production should grow steadily. The plan calls for long-term initiatives to increase oil and gas reserves. Coal-to-oil and coal-to-gas bases remain designated key areas for enhancing security capabilities. Alongside that sits a measure to replace coal-fired boilers equivalent to 30 million tonnes annually and to retrofit coal plants for co-firing with biomass or green ammonia. The plan is simultaneously decarbonising the margin and hardening the core.

Project 54Capacity additions at a fifteen year high while the output of that capacity falls. In the first year of this plan, construction and generation are telling different stories on purpose.

Coal generation down, coal capacity up, solar down 61 per cent

The operating data for 2026 complicates every simple reading of the plan. On the generation side the direction is clean. The Centre for Research on Energy and Clean Air reported thermal power generation down 4 per cent year on year in August 2026, with coal power generation down 5.2 per cent, the second consecutive monthly decline. Solar generation rose 17.1 per cent and wind 6.9 per cent. CREA's assessment was that solar and wind covered all growth in electricity demand in that month.

On the capacity side the direction inverts. Thermal additions, overwhelmingly coal, reached 50.3 GW in the first seven months of 2026, up 20 per cent year on year and the highest in fifteen years. Solar additions fell to 86.1 GW, down 61 per cent on the same period of 2025. Wind fell 12 per cent to 47.1 GW. Nuclear added 3.6 GW against nothing in the equivalent 2025 period.

These are not contradictory once the function of the capacity is separated from its output. Coal plants are increasingly being built as adequacy and flexibility assets rather than as energy producers. Their capacity factor is designed to fall. That is consistent with a system adding very large volumes of variable renewables and needing firm backup, and it is consistent with provincial incentives that reward construction. It is not consistent with the framing that coal build-out implies rising coal burn, which the generation data directly refutes.

The solar collapse needs more care. The most widely cited explanation is the pull forward of installations into the first half of 2025 ahead of the move to market based pricing for renewable output, which left 2026 comparing against an inflated base. That attribution is an assessment rather than a published official finding and should be stated as such. What is not in doubt is the magnitude. A 61 per cent fall in the world's largest solar market is a material change to global module demand, and it lands in the first year of a plan whose non-fossil target depends on sustained additions.

Security first, because the last two shocks were both supply shocks

The structure of the plan follows from what Chinese energy policy has actually been tested by. The 2021 power rationing episode and the 2022 global price shock were both failures of supply adequacy, not failures of decarbonisation. A system that has been embarrassed twice by shortage will build redundancy, and redundancy in a coal based system means coal plants that mostly sit idle.

The same instinct explains the hydrocarbon floor. Holding crude production at around 200 million tonnes a year is not an economic decision at the margin, because a material share of Chinese domestic production is high cost against imported barrels. It is an insurance decision. So is maintaining coal-to-oil and coal-to-gas capacity as designated security assets, a technology that is expensive and carbon intensive and exists primarily because it converts a domestic resource into liquid fuels without a sea lane. The stockpiling behaviour we have covered at length sits in the same logic, and the storage build-out continues regardless of demand.

Set against that, the demand side numbers suggest the insurance is being bought into a weakening market. Refinery throughput down 6.9 per cent, crude steel down 3.7 per cent and cement down 11.7 per cent describe an industrial economy consuming less energy per unit of activity and less activity overall in the heavy sectors. The property correction is the dominant cause of the cement and steel figures. The refinery number is the one with the clearest read-through to oil demand, and it is the reason forecasts of Chinese oil demand have been revised down repeatedly, a divergence we have examined between the IEA and OPEC.

The removal of the energy intensity target fits here too. If total energy demand must be allowed to grow to support electrification, AI compute load and the electrification of transport, then a binding constraint on energy per unit of GDP becomes an obstacle rather than a discipline. The plan chose growth headroom and kept the carbon metric. That is a deliberate trade and it is legible.

The peak arrives early, the plan does not claim it

The most defensible forward reading is that China's carbon emissions have already peaked on a structural basis, several years ahead of the 2030 commitment, and that the plan is written so as not to say so. Twenty one months of flat or falling emissions, with solar and wind covering all demand growth in recent months, is the signature of a system that has crossed over. The plan's language, promoting the peaking of coal and oil consumption with no date, preserves the option to be wrong without having to retract anything.

Three things would falsify that reading, and each is worth monitoring. A sustained recovery in heavy industry would restore coal burn quickly, because the idle capacity now exists to serve it. A second year of collapsed solar additions would stall the non-fossil share short of 25 per cent and force either higher coal utilisation or a visible target miss. And an acceleration in electricity demand from data centres beyond current projections would widen the gap that renewables must cover, in a period when additions are falling rather than rising.

For oil specifically the plan gives a stable supply picture and a deteriorating demand picture. Production held near 200 million tonnes, continued reserve additions and continued strategic storage construction on the supply side. Falling refinery throughput and a petrochemical pivot that converts barrels to molecules rather than fuels on the demand side, which we have covered in the PetroChina and Sinopec shift. The expansion of the mandatory carbon market into petrochemicals, already underway on the pricing timetable we examined for 2027, adds a cost layer to that pivot.

The honest summary is that the 15th Five-Year Plan is a security document with a climate chapter rather than a climate document with a security chapter, and that the first year of data suggests the climate outcome will beat the climate text anyway.

Sell into the adequacy budget, not the headline transition

The commercial misreading to avoid is treating the plan's climate targets as the demand signal. The money in the first year of this plan has gone into thermal capacity at a fifteen year high, into nuclear restart, into grid and storage to absorb variable output, and into oil and gas reserve replacement. Those are the budgets that are growing. A vendor positioning purely against the decarbonisation narrative is aiming at the slower moving line.

Three practical implications follow. First, flexibility and adequacy are the live procurement categories: plant able to cycle, grid equipment, storage, and the instrumentation and controls that let a coal unit run at low capacity factor without destroying its maintenance economics. Second, the reserve and storage programme continues independent of demand, which keeps tankage, pipework, inspection and commissioning demand stable even as throughput falls. Third, the petrochemical conversion of the refining fleet is where the refining capex is going, so proposals framed around fuels throughput are addressing a shrinking base.

On the marketing side, the discipline is the same one we apply to any policy document. Quote the plan's actual text and date rather than a secondary summary, separate the binding targets from the qualitative commitments, and state where an attribution is your assessment rather than a published finding. The buyers in this market read the primary documents. A claim about Chinese policy that cannot survive contact with the plan text costs more credibility than it buys attention.