The two-part tariff: Vietnam's capacity + energy pricing reform

What is the two-part tariff and how does it differ from today's electricity bill?

Direct answer. The two-part tariff replaces Vietnam's single-part (energy-only) retail bill with a two-component bill: a fixed capacity charge based on registered maximum demand, plus a variable energy charge billed by time-of-use band. The total bill has the same underlying cost recovery — but it is redistributed to reward flat load profiles and penalise peaky consumption.

Under today's tariff structure (Decision 1279/QĐ-BCT), a 22 kV manufacturing customer pays a single VND/kWh rate that varies by time-of-use band: peak 3,398 VND/kWh, standard 1,833 VND/kWh, off-peak 1,190 VND/kWh. Under the two-part tariff, that customer pays two separate charges each month. The first is a capacity charge calculated as registered maximum demand (in kilowatts) multiplied by a fixed VND/kW/month rate — for 22 kV manufacturing, VND 235,414/kW/month. The second is an energy charge billed by TOU band, at new rates roughly 30–35% below today's energy-only rates, because part of grid-cost recovery has moved from the per-kWh charge to the per-kW capacity charge. The customer pays the capacity fee every month regardless of consumption; if they shut down for a month, the capacity bill is unchanged, but the energy bill drops to zero.

In practice: the customer most exposed to the new structure is the one with a short, spiky peak — a factory that runs three shifts one week per month but is otherwise idle, or a data centre with large redundant-capacity reservations. The customer most rewarded is the one running a flat, high-utilisation load. BESS peak-shaving sits directly on top of this: every kW shaved from registered maximum demand cuts the monthly capacity bill permanently.

Who is in the pilot and who is excluded?

Direct answer. The pilot covers approximately 7,000 manufacturing customers consuming an average of ≥200,000 kWh per month at voltage ≥22 kV — the same eligibility universe as Decree 57's grid-connected DPPA. Hospitality, commercial, retail, and sub-22-kV customers are excluded from the pilot.

The 200,000 kWh/month threshold excludes most hotels, restaurants, and SMEs. It includes mid-to-large factories and every hyperscale data centre. EVN has publicly confirmed the pilot cohort figure; MOIT's implementation guidance extends the threshold logic by requiring TOU-capable metering for all producers and businesses with transformers ≥25 kVA or monthly consumption ≥2,000 kWh under Circular 60/2025/TT-BCT — a separate but parallel workstream that prepares the broader customer base for eventual rollout. A customer just below the pilot threshold today (say, 180,000 kWh/month) will not enter real billing in July 2026 but will enter the nationwide expansion from 2028. Customers connected at under 22 kV fall outside the pilot regardless of consumption.

In practice: for most Arcus prospects, the question is not "am I in the pilot" but "when does this apply to me." Factories above ~2 MW connected load at 22 kV+ are typically in. Every commercial-scale data centre is in from day one. Hotels and commercial sites are out of the pilot but should model the structure now — they will enter during the 2028–2030 rollout and the commercial tariff has the widest peak/off-peak spread in Vietnam, making the economics most sensitive there.

How does the capacity charge work in practice?

Direct answer. The customer declares a registered maximum demand figure (Pmax) in kilowatts when entering the pilot. Each month, the customer is billed Pmax × the VND/kW/month rate for their voltage band, regardless of whether they actually hit that peak during the month.

Pmax is measured in kilowatts over a rolling window — typically a 15- or 30-minute sliding average at the high-tension meter. The customer agrees a declared Pmax figure with their EVN regional corporation when joining the pilot; if the customer exceeds declared Pmax in a given month, penalty charges apply. If the customer consistently undershoots, they can renegotiate Pmax downward at agreed review points. The capacity charge rate by voltage band, from EVN's pilot documentation:

The structural logic is cost-reflective: higher-voltage connections use more proximate grid infrastructure and fewer downstream transformation stages, so their capacity charge is lower per kW. Sub-6 kV connections at the tail of the distribution system bear the highest per-kW capacity cost.

In practice: the capacity charge creates a strong incentive to shave peaks and declare a lower Pmax. A factory registering Pmax at 2,000 kW at 22 kV pays VND 471 million per month in capacity charges alone (USD ≈18,500 at April 2026 FX). Cutting Pmax to 1,500 kW through BESS peak-shaving saves VND 118 million per month — approximately USD 4,600 per month, USD 55,000 per year, for every 500 kW of Pmax reduction.

What happens to the energy charge?

Direct answer. Under the pilot, the energy charge falls by approximately 30–35% versus today's Decision 1279 rates — because part of grid-cost recovery has moved to the new capacity component. The TOU band structure (peak, standard, off-peak) is retained.

EVN's pilot documentation shows draft pilot energy rates of VND 843–904/kWh off-peak, VND 1,253–1,332/kWh standard, and VND 2,162–2,251/kWh peak, with the exact number within each range depending on voltage band. Compared with Decision 1279's 22 kV manufacturing rates of VND 1,190 off-peak, VND 1,833 standard, VND 3,398 peak, the pilot rates represent a ~25–35% reduction in the variable bill component. The peak–off-peak arbitrage spread therefore narrows from VND 2,208/kWh under today's structure to roughly VND 1,300–1,400/kWh under the pilot — a compression of ~35–40%.

In practice: this is the single most important structural effect for BESS business cases. The energy arbitrage revenue (peak minus off-peak, adjusted for round-trip efficiency) falls by roughly a third under the pilot. But that revenue loss is more than offset by the capacity-charge saving from peak-shaving, provided the customer has a peaky load that can actually be shaved. A flat baseload data centre gains little from peak-shaving; a batch-manufacturing factory with sharp demand spikes can gain substantially more than it loses on arbitrage compression.

How does this interact with DPPA and the new TOU windows?

Direct answer. Decree 57 DPPA offtakers face a potential double-charging risk because both the new capacity charge and the CDPPA grid service charge under Decree 57 recover grid-infrastructure costs. MOIT's separate April 2026 consultation on new TOU windows will reshape when capacity-reducing peak-shaving is most valuable.

Norton Rose Fulbright confirmed in January 2026 commentary that DPPA offtakers under Decree 57 Model 2 (grid-connected) will likely still pay the full capacity charge to EVN — the CDPPA charge recovers similar grid costs through a different mechanism, but there is no netting specified in either regulation. An amending decree to Decree 57 is expected in 2026 to resolve the overlap; until then, any DPPA term sheet signed under the pilot cohort should include a change-in-law clause specifically addressing the double-charging risk. Separately, MOIT published a public consultation draft on 15 April 2026 proposing new TOU windows under Circular 60/2025/TT-BCT — both the MOIT seasonal proposal and EVN's counter-proposal eliminate the 09:30–11:30 morning peak. If the EVN proposal (single 17:30–22:30 evening peak) is adopted, BESS peak-shaving becomes more valuable because the current split peak (2 hours morning + 3 hours evening) consolidates into a continuous 5-hour evening block — the total peak hours stay the same, but a single longer discharge cycle is more efficient than two shorter ones and removes the morning ramp-down cost.

In practice: any project modelling two-part tariff economics in 2026 should do so against three scenarios — current TOU, MOIT seasonal proposal, EVN evening-only proposal. The hybrid solar + BESS configuration wins under all three, but the exact payback delta shifts with which TOU regime is adopted. Watch for the MOIT final decision in Q3 2026.