Malaysia’s Data Centre Boom: The Signal Investors Should Test
The signal: power, water and who pays for them may now decide who profits from Malaysia’s data centre growth.
Why now: Budget 2027 is scheduled to be tabled on Friday, 9 October. The government has already said data centre incentives will be tied to energy and water efficiency.
Quick Answer
Malaysia’s government has said data centre incentives will be tied to energy and water efficiency, and Moody’s projects data centres could use 31% of Peninsular Malaysia’s electricity by 2035, up from 6% in the first half of 2026. For investors, the useful question becomes: who benefits when power and water are the constraint?
I’ve spent 30 years around data, customers and growth plans. One pattern keeps repeating: when demand runs ahead, the bottleneck decides who makes money.
Most coverage of Malaysia’s data centre boom has focused on demand. The latest signals point to supply: electricity, water and who carries the cost. That affects which businesses are worth watching.
What happened
The policy line. In its Pre-Budget Statement on 18 August 2026, the Ministry of Finance wrote: “Data-centre incentives will be tied to energy and water efficiency so that digital growth does not strain supplies or raise costs for the people.” (MOF)
The demand numbers. Moody’s, as reported on 17 September 2026, estimates data centres used about 4% of Peninsular Malaysia’s electricity in 2025 and 6% in the first half of 2026, and projects 31% by 2035. Actual data centre load reached 1.26 GW in June 2026. Moody’s estimates RM80 billion to RM95 billion of generation investment is needed over ten years, while 13 GW of existing coal and gas capacity expires between 2026 and 2035. (TNGlobal, citing Moody’s)
Who pays. In December 2025 the government told Parliament it follows a “user pays” approach: data centre operators bear the extra cost of strengthening the grid, and domestic water tariffs will not be raised just to serve data centres. The same statement put committed demand under electricity supply agreements with TNB at over 7,000 MW as of September 2025. (Bernama)
Notice the gap. Over 7,000 MW committed (September 2025). About 1.26 GW of actual load (June 2026). These are different measures on different dates, so don’t compare them as a ratio. But the distance between “contracted” and “running” is exactly where investor assumptions can drift away from reality.
Why it may matter
When demand outruns supply, value tends to move to whoever can add capacity. The cost of that expansion has to land on someone.
Three shifts are worth testing.
1. Efficiency becomes a cost advantage
If incentives depend on energy and water efficiency, operators with efficient designs may get better terms. Less efficient projects may face higher costs or slower approvals. This is an inference: the Budget detail will show how efficiency is measured.
2. Value may move along the value chain
Land, construction and fit-out benefit first. Generation, grid upgrades, cooling, water treatment and recycling become more important as supply tightens. A business tied only to the first wave carries a different risk from one tied to the second.
3. “User pays” moves cost onto operators
If operators fund grid upgrades and pay a separate tariff, their returns depend on passing those costs to hyperscaler customers. Revenue can grow while margins come under pressure.
Learn to read signals like this with Claude
At the Claude Smart Investors Workshop you’ll build a routine that turns news like this into a clear check on what you own. Saturday, 14 November 2026, 9am to 5pm.
See the workshop detailsWhat to watch
| Watch | Why it matters |
|---|---|
| Budget 2027 detail on data centre incentives (9 October) | How “efficient” is defined, and whether it applies to new projects only |
| Committed demand turning into actual load | Contracted power is not revenue until it runs |
| Grid and generation capex announcements | Shows who funds the buildout, and on what terms |
| Data centre tariff and cost-recovery decisions | Decides whether operators or consumers carry the cost |
| Hyperscaler capex guidance | Moody’s flags investment deferral as a risk if AI demand fluctuates |
The open question: can supply keep up with contracted demand without squeezing margins somewhere in the chain?
What would weaken the signal
- Budget 2027 keeps incentives broad, with no meaningful efficiency test.
- Committed demand converts to actual load smoothly, with no visible strain on reserve margins.
- Operators pass grid and tariff costs to customers without any margin pressure.
If you see these, the simpler “more data centres, more demand” story still holds.
Test it against your holdings
Copy this into Claude. If you’ve saved your holdings and thesis in a Claude Project, run it there.
Prompt: Data Centre Signal Check
ROLE You are my investment research partner. I make the decisions. You test signals against what I own. PROBLEM Signal to test: Malaysia's data centre growth may be shifting from a demand story to a supply story. Incentives are being tied to energy and water efficiency, power demand is rising fast, and grid costs are moving onto operators. Value may shift towards power, efficiency and water, and away from businesses tied only to construction and land. INSTRUCTIONS 1. If you already have my holdings and thesis, use them. If not, ask me for my holdings and stop until I reply. Ask for any thesis assumptions you need. 2. For each holding, identify its role: operator, builder, landlord, power or grid, equipment or cooling, water, customer of data centres, or none. 3. Look for evidence of real revenue from data centres (recognised revenue, not MOUs, tenders or "pipeline"). 4. Assess whether efficiency-linked incentives or "user pays" grid costs could help or hurt its margins. 5. Name the valuation assumption that would change, and what evidence would confirm or weaken the signal. GROUND RULES - Separate announced plans from measured results. - Cite a source for every factual claim. Mark unconfirmed items [CHECK]. - If a link to the signal is weak or indirect, say so. - No trade recommendations. Don't change any of my files. STRUCTURE One short block per holding, ending with one verdict: WATCH, DIG DEEPER, or NOT RELEVANT. MILESTONE In five minutes I know which holdings this signal touches, and what to check after Budget 2027 is tabled.
The right answer is often NOT RELEVANT. That’s useful too: it tells you what you can stop worrying about.
Running this may change nothing in your portfolio. You’ll still know which holdings deserve a closer look after 9 October.
FAQ
What did the government say about data centres before Budget 2027?
In the Pre-Budget Statement released on 18 August 2026, the Ministry of Finance said data centre incentives will be tied to energy and water efficiency so that digital growth does not strain supplies or raise costs for the people. How efficiency will be measured had not been published at the time of writing.
How much electricity do data centres use in Malaysia?
According to Moody’s, as reported on 17 September 2026, data centres accounted for about 4% of power consumption in Peninsular Malaysia in 2025 and 6% in the first half of 2026. Moody’s projects this could reach 31% by 2035.
Is the data centre boom good for Bursa Malaysia stocks?
It depends on the company’s role. Builders, power suppliers, equipment makers, landlords and operators are exposed in different ways, and efficiency-linked incentives could change who benefits. Test each holding’s specific exposure rather than treating data centres as one theme.
What is the difference between committed and actual data centre demand?
Committed demand is power that operators have contracted to take in future. Actual load is what is being used today. Contracts can be delayed, resized or cancelled, so revenue assumptions should follow the conversion of committed demand into real load.
Can I use Claude to check whether my holdings are exposed?
Yes. The prompt in this post asks Claude to map each holding’s role in the data centre value chain, find evidence of real revenue, and end with WATCH, DIG DEEPER or NOT RELEVANT. Claude analyses; it does not recommend trades.
Test the headlines against what you own
The Claude Smart Investors Workshop shows you how to set up Claude as your research partner: write your thesis, monitor what you own, and run signal checks like this one in minutes.
Saturday, 14 November 2026 · 9am to 5pm
Reserve your seat Get the next Signal Watch in the free communityThis brief is a point-in-time educational analysis based on public sources available on 3 October 2026. It is not investment advice or a recommendation to buy, sell or hold any security. Figures are as reported by the cited sources. Do your own research and consider speaking to a licensed adviser before making investment decisions.
Jane Chew / AI Strategy Coach, HRDC Corporate Trainer / Founder of digitalaibusinessclub.com