Investment Committee Primer — No Power-Sector Background Required

30-second elevator pitch: The Southeast Asia data center wave (Microsoft, Google, AWS, ByteDance, Tencent) is landing in Johor Bahru, adding 1–2 GW of new demand each year that the TNB grid bottleneck cannot absorb. We install rooftop PV and batteries at industrial sites, consume on-site, sell to the data centers next door, and export the surplus clean power to Singapore — one asset, three revenue streams: avoided cost (the electricity bill we no longer pay) + a “clean-power premium” paid by hyperscalers + cross-border export uplift.

Current finding: 60 MW of PV alone clears a 13% IRR — bankable today. BESS has to wait for the 2028+ capex curve to make money. Carbon value is the biggest single lever — at the hyperscaler internal carbon price (USD 100/tCO₂), the “clean-power premium” alone is worth 2.5× the avoided-cost layer.

You only need to remember four numbers

Number	What it is	The reaction you should have
+82.7 M RM	NPV of the 60 MW PV single-site case, 13% IRR, 8-year payback	”PV clears today and passes a project-finance bar.”
−349 M RM	NPV gap of adding BESS at the same site	”BESS is loss-making if forced today — wait.”
+211 M RM	Additional NPV from the hyperscaler carbon stack	”The real money is selling certificates, not selling kWh.”
+570 M RM	Structural estimate of NPV from a 100 MW cross-border allocation	”Securing the ENEGEM allocation is the single biggest lever.”

For the full numbers see headline findings or the executive summary.

The business model in plain terms (no power-sector vocabulary)

Think of this project as a multi-revenue asset — one piece of land, one stack of capital, sold simultaneously to three different buyers:

Revenue stream 1: Avoided cost (the electricity bill we no longer pay)

The factory installs PV, and every kWh generated during the day is one less kWh purchased from TNB. TNB’s industrial tariff is time-of-use: the peak window from 14:00 to 22:00 is 67% more expensive than off-peak — and the sun is strongest precisely during that peak window, so the savings are largest.

Analogy: like rooftop solar on an apartment building, the saving is your monthly electricity bill — except the factory is 1,000× larger, and so is the bill.

This is a cashflow you can compute today — it doesn’t depend on any new contract or new policy.

Revenue stream 2: Clean-power certificates (carbon attribute)

Microsoft, Google, and the rest of the hyperscaler cohort have published internal carbon prices in the USD 100/tCO₂ range (Microsoft USD 100, Google USD 25, Meta USD 64 are publicly disclosed). The interpretation: they are willing to pay USD 100 to neutralise a tonne of CO₂ they emit.

If you sell them certified clean power (each MWh paired with a certificate proving zero carbon), they will pay a premium above the standard tariff. That premium is called the carbon attribute premium.

Analogy: like organic vegetables — a head of lettuce is worth RM 5 on its own, but with an “organic certified” label it sells for RM 15. The RM 10 difference is the attribute.

The math: MY grid average carbon intensity 631 gCO₂/kWh × USD 100/tCO₂ × 4.6 RM/USD ≈ RM 290/MWh of premium. That’s 86% of TNB’s industrial tariff of RM 338/MWh. Effectively a doubling of the price.

This is the biggest gold mine of the next 5–10 years — but it depends on (a) hyperscalers actually paying, and (b) us obtaining the certification.

Revenue stream 3: Cross-border export to Singapore (ENEGEM)

Singapore generates 90% of its electricity from gas and pays roughly 3× Malaysia’s tariffs. The two governments signed an agreement allowing Malaysia to export up to 1 GW of clean power to Singapore (auctioned in tranches, ongoing).

Whoever wins the export allocation gets to sell daytime PV surplus into Singapore at Singapore prices. Singapore wholesale (USEP) trades at USD 80–200/MWh; net of cross-border wheeling, losses, and intermediary margin, this lands at RM 143–510/MWh.

Analogy: the SGD/MYR FX gap is roughly 3.4×, so cross-border power is essentially the daigou (cross-border resale) trade — wholesale into Singapore, retail to SP Group.

The math: 100 MW allocation × 25% capacity factor × 20 years ≈ +570 M RM NPV (base USEP 120 SGD/MWh).

⚠️ Note: this is a structural estimate, not a backtest — we don’t have an EMC USEP historical time series, so this is an upper bound, not an expected value. Whether we win the allocation is a commercial / regulatory question, not a technical one.

Revenue stream 4: Grid services (VPP service)

The grid is sometimes stressed (summer heat, plant outages) and needs instantaneous load adjustment. If our batteries can dispatch on grid command, the grid pays a capacity charge plus a dispatch fee.

Analogy: like the standby diesel gensets in a large hotel — idle most of the time, but ready to fire when the grid needs them. The difference is that our “genset” is a battery — fast, distributed.

Malaysia’s VPP service market hasn’t crystallised yet (ST opened consultation in 2024); our model assumes 2027–2030 capacity payments of 60–120 kRM/MW/month. This is option value, not contracted cashflow.

Glossary (mapped to concepts you already understand)

Term you’ll see	Concept you already understand	One-line explanation
BTM (behind-the-meter)	On-site / internal consumption	PV sits behind the factory’s meter, so generation directly offsets purchased energy without going through the grid market
ETOU	Time-of-use tariff	TNB industrial tariff with peak/off-peak split: peak 0.337 RM/kWh, off-peak 0.202, a 67% spread
ICPT	Fuel cost pass-through	TNB’s semi-annual fuel surcharge, currently +16 sen/kWh; could be reformed at the next review
MD (Maximum Demand)	Peak-demand charge	Factory’s monthly peak instantaneous power × RM 35/kW/month — this is the line item BESS primarily reduces
BESS	Industrial battery system	LFP (lithium iron phosphate) chemistry, 1 MWh in a 40-ft container, ~6,000 cycle life
PV	Solar photovoltaics	Rooftop or ground-mount array, performance ratio (PR) 0.78 (industry benchmark)
PUE	Data center efficiency ratio	Total power / IT power, 1.4 means 40% goes to cooling and overhead
VPP	Virtual power plant	A dispatchable “plant” assembled from many distributed batteries / PV
DRC / Demand Response	Demand-response contract	Curtail load on grid signal, paid per kW × tariff
24/7 CFE	24/7 hourly clean-energy matching	The RE100 upgrade — 100% clean every hour, not just netted annually
REC	Renewable Energy Certificate	1 MWh of clean energy = 1 certificate, tradable separately (common in the US)
NEM 3.0	Net energy metering scheme	MY’s current export-billing policy
ENEGEM	Malaysia’s energy exchange	Cross-border export auction platform
USEP	Singapore wholesale price	Spot price settled every 30 minutes by EMC (Singapore’s energy market company)
LP / linear programming	LP solver	The math that solves for optimal BESS charge/discharge schedules (we use CBC)
NPV / IRR / payback	You already know	Project finance trio; the model uses 8% discount rate, 20-year tenor

Why PV is doable today and BESS isn’t

The arithmetic in one paragraph:

PV today: 1 MW × RM 3.5 M/MW capex → RM 532k/yr saved on the bill → simple payback 6.5 years, 13% IRR. Clears a project-finance bar.

BESS today: 1 MWh × RM 5.0 M/MWh capex → RM 100k/yr saved (peak/off-peak arbitrage + MD reduction) → simple payback 50 years. Nowhere near bankable.

Where’s the gap?

PV unit price    RM 3.5M/MW    Annual return   RM 532k/MW    IRR  13%   ✅ investable
BESS unit price  RM 5.0M/MWh   Annual return   RM 100k/MWh   IRR  −5%   ❌ not investable

BESS flips only when two things happen together:

1. Capex falls to ≤ 2.0 M RM/MWh: BNEF projects this around 2028–2030. LFP shipments quadruple from 2024–2030, prices roughly halve.
2. A VPP service contract is signed: ≥ 60 kRM/MW/month (the “Strong VPP stack” tier). This depends on when MY rolls out DRC / capacity payments.

Our model triggers in 2027 (if both conditions land). From an IC perspective:

• 2026–2027: PV only, no BESS.
• 2027–2028: small BESS pilot (1 MWh scale), accumulate operating data and relationships.
• 2028+: BESS at scale.

Detailed analysis: BESS investment trigger curve.

Three uncertainties the IC should track

Uncertainty 1: ICPT policy reform (regulatory risk)

ICPT is TNB’s semi-annual “fuel surcharge”, currently +16 sen/kWh. It accounts for a meaningful slice of NPV — if reform removes or sharply caps it, the PV-alone case drops from +83 M to +5 M NPV.

Mitigation: write a tariff-floor clause into the PPA, locking the ICPT risk on the customer side. Monte Carlo shows this clause drops PV-alone P(NPV<0) from 12.6% to 0%. See risk-adjusted NPV.

Uncertainty 2: BESS capex curve (technology risk)

We assume BESS capex hits RM 2.0 M/MWh by 2028. If the BNEF curve slips 1–2 years, the BESS-bankable window slips with it.

Mitigation: the PV-only case doesn’t depend on the BESS capex curve. Build PV first; commit to BESS only when the capex actually arrives.

Uncertainty 3: Hyperscaler build-out pace (market risk)

If JB data center construction is delayed (geopolitics, regulatory tightening, demand allocation caps), the anchor-tenant base shrinks.

Mitigation: (a) even without hyperscalers, PV BTM still works at ordinary factories (carbon stack is smaller), and (b) the stylised composite portfolio is already diversified across 10 sites — losing any single anchor is not fatal.

Detailed risk register: risk-adjusted NPV.

How we plan to validate next

Next IC step:

We plan to find a Johor Bahru factory and build a 1 MW pilot plant, with four objectives:

1. Cashflow validation: 12 months of measured operating data vs LP-model forecast, errors within ±10%
2. TNB process validation: run the NEM 3.0 application, grid-interconnect approval, and ICPT settlement end-to-end — produce a playbook for downstream replication
3. Relationship building: establish working relationships with EPC, TNB, ST, SEDA, IRDA
4. Cost-curve lock-in: measured EPC turnkey pricing, O&M cost, performance guarantee terms — turning the financial model from theoretical numbers into cost data

Budget: RM 5–10 M (PV plus a small BESS configuration).

Timeline: 2026 Q3 kickoff → 2027 Q1 grid-connection → 2028 Q1 validation report.

Detailed plan: 1 MW pilot plant

One-page summary (to take back to your fellow GPs)

1. JB data center wave → TNB grid bottleneck → window for BTM self-consumption + cross-border export
2. PV is investable today (+83 M / 13% IRR / 8-year payback); BESS waits for the 2028+ capex curve
3. The real money is in the carbon attribute (+211 M) and the ENEGEM cross-border allocation (+570 M), not in the kWh
4. Three risks: ICPT reform, BESS price curve, hyperscaler pace — each has a contractual mitigation
5. Next step: 1 MW pilot validation, RM 5–10 M budget, measured data within 12 months

If anything still feels unclear, go straight to the executive summary (5-minute read, full headline numbers + risk register + recommended actions).