CAAICE
25 July 2026

Twenty-one years, seven reviews

Malaysia's CRESS asks corporate buyers to sign twenty-one-year contracts around a charge that can be reset every three years. The allocation of that risk has not been explained.

DayOne, a hyperscale data centre developer, has contracted for the output of 1.5 GW of solar generation and 2.2 GWh of battery storage from subsidiaries of Tenaga Nasional Berhad. The agreements run through Malaysia’s Corporate Renewable Energy Supply Scheme. They are among the largest corporate renewable procurements ever signed in Southeast Asia.

Tenaga has now signed agreements covering forty-nine data centre projects. Together those projects could draw as much as 7 GW. Announced CRESS project capacity stood at roughly 4 GW at the start of this year.

These are serious numbers, arrived at quickly, under a scheme that has existed since September 2024.

What CRESS does

Before CRESS, a Malaysian company that wanted clean electricity had a short list of options and none of them worked at scale. It could put solar on its own roof. It could buy a green tariff from the utility. It could purchase renewable energy certificates and leave the physical electron question unanswered.

CRESS added the missing option: buy directly from an independent renewable generator, and use the grid to get the power from there to here. Open access, in the language the sector uses. The generator and the buyer negotiate a price between themselves. The grid operator is paid separately for transporting the electricity.

That separate payment is the System Access Charge, and it is where the difficulty sits.

The charge

The SAC is levied per kilowatt-hour delivered. It is differentiated by the character of the supply. Generation that cannot guarantee output when the grid needs it pays more; generation that can pays less.

At launch the non-firm rate was set at 45 sen/kWh and the firm rate at 25 sen/kWh. Following revisions, the figures have come down, with the non-firm rate now in the region of 40 sen/kWh and the firm rate at 20 sen/kWh. Wood Mackenzie has put the non-firm charge at roughly US$90/MWh and described it as the highest system charge applied to corporate PPAs anywhere in the region.

Practitioners should confirm current figures against the prevailing guidelines rather than relying on this note. The rates have already been revised once and the direction of travel is downward.

The level is not the problem this brief is concerned with. A high charge is a commercial fact. Parties can price it, argue about it, or walk away from it.

The problem is what happens to it over time.

Seven reviews

CRESS contracts run for twenty-one years.

The SAC is reviewed every three years. Each review may move the charge by up to fifteen per cent from the prevailing level.

Twenty-one years divided by three gives seven reviews across the life of the contract. Fifteen per cent compounds. Seven consecutive increases at the cap would take a charge of 40 sen/kWh to roughly 106 sen/kWh. Seven consecutive decreases would take it to about 15 sen/kWh. Neither extreme is likely. Both are permitted.

That range is wider than the margin on most of these projects.

A buyer signing a CRESS PPA today is therefore agreeing to a delivered cost of power in which one large component is known for three years and, after that, sits inside a band the parties cannot close.

Why this is a drafting problem

Analysts have been calling the SAC opaque and expensive for some time. The observation is correct and it has been made well by others, including in a joint intervention from SEMI and Wood Mackenzie arguing for published methodology and clearer cost components.

The question that follows from it has been asked less often. In a twenty-one-year agreement, who bears the review risk, and through what mechanism?

There are only three answers.

The buyer can absorb it. The generator sells at a fixed strike price and the buyer takes the delivered cost as it comes. This is clean and it is what most buyers say they will not accept, because the whole reason a corporate signs a twenty-one-year PPA is to remove electricity cost volatility from its books. A contract that removes generation price risk while leaving an uncapped transport charge floating has solved half the problem and charged the buyer for the privilege.

The generator can absorb it. The parties agree a delivered price and the generator wears the SAC movement. Generators will do this, at a price. The premium they charge reflects a risk they cannot hedge, cannot model with any confidence, and cannot lay off to anyone else. That premium is paid on day one against a risk that may never materialise. It is an expensive way to buy certainty.

Or the parties can share it, which in practice means writing a change-in-law or change-in-cost provision and hoping it holds. This is where most of these agreements end up, and it is where most of the future disputes will come from.

The reason is straightforward. Change-in-law clauses are drafted to handle events. A regulator does something unexpected; the clause allocates the consequence. A scheduled triennial review conducted under published rules is not an unexpected event. It is a term of the contract. A clause that treats it as a supervening event is being asked to do work it was not designed for, and a tribunal reading it fifteen years from now will say so.

Drafting around a known, scheduled, bounded variable requires a different instrument: a pass-through with a defined base, an index, a collar, a reopener with agreed consequences if the parties fail to agree. These are ordinary tools. They are used in gas contracts, in tolling agreements, in availability-based infrastructure concessions. They are not, so far as public reporting shows, being used consistently here.

The transparency point is a pricing point

The SEMI and Wood Mackenzie argument for transparency is usually framed as a matter of good regulatory practice. It is that. It is also a commercial matter with a number attached.

An index only works if the parties can observe what it is indexed to. A collar only works if both sides can form a view on where the underlying is likely to sit. If the methodology behind the SAC is not published, a generator asked to take review risk cannot price it except by guessing high, and a buyer asked to take it cannot budget for it except by assuming the worst.

The opacity is therefore not merely a governance concern. It is a direct input into the strike price of every CRESS PPA signed while it persists, and it is paid for by whichever party ends up holding the exposure. That cost is invisible because it never appears as a line item. It is embedded in a strike price that is a few ringgit higher than it needed to be.

Publishing the methodology would reduce the delivered cost of renewable power in Malaysia without changing the charge by a single sen.

What this has to do with the grid

There is a reason to think the review mechanism will be tested rather than left alone.

In February the Prime Minister said the government had spent the previous eighteen months to two years restricting non-AI data centre development. Johor has rejected close to a third of the applications it has received, with energy efficiency among the grounds. Demand for grid capacity is running well ahead of what the system can absorb.

The SAC is one of the few instruments available to manage that. It is a per-kilowatt-hour charge, differentiated by firmness, adjustable on a three-year cycle. If policy wants to push new load toward firm supply, or to recover transmission investment from the users driving it, the SAC is the lever that is already in place and already scheduled to move.

Which is a reasonable thing for a regulator to do. It is also precisely why buyers should not assume the charge will stay where it is.

What would help

Three things, in order of how easily they could be done.

Publish the methodology. Not the number, which is already public, but the calculation: what costs are recovered, on what basis they are allocated between firm and non-firm supply, and what would cause them to move. This is the single change that would most reduce the cost of capital for CRESS projects.

Publish the review record. When a review concludes, publish the inputs and the reasoning alongside the result. Three cycles of this and the market would have something to index against. At present there is nothing.

Consider a longer horizon for the tail. A charge that is reviewable every three years for twenty-one years is a different instrument from one that is reviewable for the first nine and fixed in real terms thereafter. If the policy objective is to attract long-term capital into generation, the tail of the contract is where certainty is worth most and where the option value of adjustment is worth least.

Beyond Malaysia

Every country in the region that opens third-party grid access will face this. The access charge is the mechanism by which a system operator is made whole for carrying power it did not generate, and no one has yet designed one that is simultaneously cost-reflective, transparent, and stable enough to underwrite a twenty-year contract.

Malaysia is further along than most. That is why its difficulties are worth reading closely rather than dismissively. The scheme has moved several gigawatts of corporate demand into contracted renewable supply in under two years, which is more than most of its neighbours have managed.

The question is whether the contracts underneath those gigawatts will hold when the first review lands.

CAAICE publishes all research free of charge. This brief draws on publicly available sources and does not constitute legal advice. Corrections and challenges are welcome at [email protected].

Sources