Malaysia Electricity Price Outlook 2027

In November 2025 the Automatic Fuel Adjustment was minus 8.91 sen/kWh. In August 2026 it is plus 3.80. That is a 12.71 sen swing on every kWh in fourteen months, on a base tariff of 45.62 sen. Anyone building a 2027 OPEX line is really asking one question about the Malaysia electricity price: which way does that swing run next, and how much of it is even forecastable?

Short answer: three of the four pressures on 2027 bills are forecastable enough to bound, one is not, and the honest output is a range, not a number. Below is the evidence for each, with a confidence level attached and opinion labelled as opinion.

Where the Malaysia electricity price sits in August 2026

Two numbers do the work. The RP4 base tariff for Peninsular Malaysia is 45.62 sen/kWh, approved by the Energy Commission for the 2025 to 2027 regulatory period, up 14.2% from RP3's 39.95 sen (TNB Bursa announcement, 26 Dec 2024). That one is fixed. The second is reset monthly.

The published AFA series since RP4 took effect: 0.00 sen in July 2025, then widening discounts through the second half of the year peaking at -8.91 in November, narrowing through -4.99 in January 2026, -2.15 in March and -0.47 in April, then flipping positive at +1.38 in May, +2.59 in June, +3.59 in July and +3.80 in August (paultan.org, 1 Aug 2026). The mechanism itself is covered in our AFA explainer; what matters for budgeting is the shape. Five straight months of movement in one direction, and a full-range spread larger than a quarter of the base tariff.

Two things that trend line hides. First, August's headline +3.80 is a post-subsidy figure: the pre-subsidy AFA was 4.28 sen/kWh, and the KWIE industry fund absorbed 11% of the cost increase, worth 0.48 sen/kWh, down from 35% and 1.96 sen/kWh in July (paultan.org, 1 Aug 2026). The cushion is real and it is shrinking. Second, the plus or minus 3 sen figure commonly described as an AFA "cap" behaves as a review threshold, not a ceiling: published rates have run to -8.91 and +3.80, with movement beyond the band understood to require cabinet approval rather than being blocked.

What four pressures will shape 2027 bills?

Fuel and FX pass-through: high confidence, and the driver is coal, not gas

The August 2026 calculation is instructive because the two fuels point opposite ways. Tier 1 piped gas at 31.71 RM/mmBTU sits below its 35 RM/mmBTU tariff-setting base, while coal at USD 131.71/MT sits well above its USD 97 base, and the ringgit at 4.0557 to the dollar amplifies the coal gap rather than offsetting it (paultan.org).

Our read: a 2027 budget built on "regional gas prices look calm, so the surcharge will fade" is watching the wrong variable. Coal and USD/MYR are carrying the surcharge right now, and neither is something Malaysian policy controls.

TNB also publishes a rolling three-month projection alongside the monthly rate. At the end of July 2026 it pointed upward through the fourth quarter rather than back toward discount territory. Projections get revised, and they should never be pasted into a budget as fact. But nothing in the published series supports assuming a return to discount territory in 2027 without a coal or currency reversal. Sites that also burn piped gas should note the separate distribution tariff reset: Gas Malaysia's base average distribution tariff is RM1.880/GJ/day for 2026 to 2028, RM2.114 after the IBR adjustment (The Edge Malaysia).

Carbon tax: low confidence of any 2027 bill impact

The carbon tax on iron, steel and energy was announced for 2026 and reaffirmed in the Budget 2026 speech in October 2025. In April 2026 the Minister for Environmental Sustainability and Natural Resources said implementation is under review "in light of current geopolitical circumstances", with the near-term focus moved to verifying carbon credits (Eco-Business, 22 Apr 2026).

Our read: do not provision a carbon-tax line on your 2027 electricity bill. Do keep it as a tracked watch item tied to the National Climate Change Bill, because a generation-sector carbon price would eventually reach C&I users through generation cost, not as a separate charge. The scope and rate detail sits in our Malaysia carbon tax article.

RP4 trajectory and the RP5 determination: high confidence for 2027, watch 2028

RP4 runs to 31 December 2027. That makes 2027 the last year of a known structure and, simultaneously, the year the next determination is negotiated and decided. Our read: 2027 is your most predictable year on the structural side and your least predictable year for forward guidance. If your planning cycle sets three-year budgets, the 2028 line is the one carrying real regulatory risk, not 2027.

Green procurement demand: medium confidence, and the effect is indirect

The Green Electricity Tariff premium has been a single tier since 1 July 2025, priced by commitment length at 5 sen/kWh for one year, 4 sen for two and 3 sen for three, against a programme quota of 6,600 GWh. On the bilateral side, CRESS has an announced project pipeline of roughly 4 GW, and the Energy Commission revised the CRESS guidelines on 29 December 2025 (Baker McKenzie, on both schemes).

Our read: neither scheme lowers the base tariff for anyone. They change who pays for grid capacity, and the variable to watch is the System Access Charge, whose trajectory is the main open question for anyone modelling a CRESS PPA against a grid counterfactual. We work the GET premium against AFA volatility in the GET 2026 hedge math.

What is not likely to change before 2028

The base tariff of 45.62 sen/kWh, the RP4 tariff structure, the split of medium-voltage demand charges into a Capacity Charge of RM 89.27/kW and a Network Charge of RM 97.06/kW per month, the 1.6% KWTBB levy on non-domestic bills, and the power factor surcharge thresholds of 0.85 below 132 kV and 0.90 at or above it. All of that is locked for the regulatory period. Budget those as constants and spend your uncertainty allowance on the AFA line alone.

How should a 2027 budget handle this?

Plan sensitivities, not point estimates. Take an illustrative site at 500,000 kWh per month. Every RM0.01/kWh of AFA movement is RM5,000 a month, RM60,000 a year. The observed range since RP4 began, -8.91 to +3.80 sen, is RM63,550 a month between the best and worst month actually published. That is the size of the exposure you are being asked to guess at, and it is illustrative only: scale it to your own kWh.

Our read on a defensible three-case frame, all clearly assumption-driven: a low case at AFA zero, a base case at the trailing twelve-month average of the published series, and a high case anchored to the top of TNB's published forward projection. Document which case each department signed off on. When the variance conversation happens in Q3 2027, the argument you want to have is about which case landed, not about whose forecast was wrong.

Lock in the controllables, because they hedge whichever way the tariff goes

Maximum demand is the one line where the arithmetic is unambiguous. At RM 89.27 plus RM 97.06 per kW per month, every kilowatt of recorded MD you permanently remove is worth RM 2,235.96 a year, and that saving is entirely independent of what the AFA does. There is no forecast in it. The tactics are in our guide to cutting TNB maximum demand charges.

Power factor is the second: staying above 0.85 below 132 kV avoids a surcharge that is pure waste, and correction hardware pays back on a schedule you can actually calculate. Load shifting is the third, and it is the one that compounds with AFA risk rather than merely sitting beside it, because moving kWh out of peak windows reduces both the energy volume exposed to the surcharge and the demand peak that sets your capacity charge.

Our read, stated plainly: you cannot forecast the 2027 Malaysia electricity price to a useful precision, and you should stop trying. You can bound it, and you can shrink the denominator it multiplies against. The second is where the return is.

If you want the MD, power factor and load-profile picture for your own site before the 2027 budget locks, request a demo and we will walk through what your interval data already shows.