Data centres next door: At what cost?

IN Section 51a of Petaling Jaya, minutes away from mature residential areas, stands an Australian-owned RM2.8bil data centre (DC) that was just switched on two months ago.

In Bukit Jalil, a sprawling multi-block, multi-storey DC complex now stands just a four-minute drive from a primary vernacular school.

In Jalan Sri Permaisuri, Cheras, two DCs have been reported as being built, one of which is a seven-storey DC. In Kota Damansara, a proposed three-storey DC in a residential area has triggered public backlash and prompted a Cabinet minister to openly oppose the project.

And the list goes on.

Malaysia’s DC boom is entering a new phase. The first wave focused on large purpose-built campuses in locations such as Johor and Cyberjaya. A second wave is increasingly moving into mature urban areas. This shift deserves greater scrutiny.

“We do not discount the possibility of further interest in urban area DCs going forward,” says RHB Research in a July 13 note.

Urban DCs are not necessarily small. For example, the 65 megawatt (MW) facility by NEXTDC in Section 51a, known as KL1, is a hyperscale DC. The first stage, with a capacity of 15MW, has been operational since May.

It was previously reported that a 100MW DC can use about 1.1 million gallons of water per day, roughly the amount used by 10,000 people. DCs require water for cooling.

Johor, which hosts the bulk of Malaysia’s DC capacity, has stopped approving Tier 1 and 2 DCs. These facilities can consume up to 50 million litres of water a day, reportedly enough to meet the daily drinking needs of 25 million people.

Greenpeace Malaysia campaigner Dunxin Weng (Yan) says Johor had previously approved DCs “too quickly” and now, there is a need for immediate regulatory tightening.

About 80% of leased DC net information technology (IT) power capacity is located in Johor, according to S&P Global Ratings.

Water demand is only one issue. Electricity consumption and noise pollution are becoming equally important considerations as DCs move closer to residential areas.

Of course, it may not be appropriate to treat all DCs alike. Efficiency matters.

NETXDC claims that rainwater harvesting at its KL1 facility reduces potable water consumption by more than 35%.

It also says KL1’s has a power usage effectiveness (PUE) of less than 1.4 at 25% IT load, placing it among the most efficient DCs operating in hot and humid climates.

RHB Research says newer DCs in Malaysia have generally begun adopting cooling methods with less water usage to address potential supply constraints. Moves to recycle water for cooling are also underway.

Policy gaps

DCs are undoubtedly critical infrastructure for cloud computing, artificial intelligence (AI), digital finance and online services. Few would dispute their importance to Malaysia’s digital ambitions.

Public discussion has largely revolved around whether Malaysia should welcome more DC investment.

However, in recent times, Malaysians have been asking a more pressing question: Should DCs be built in urban areas closer to homes?

Urban sites clearly offer strategic advantages, enough to justify the premium developers pay for city land instead of building farther away from major population and business centres.

Klang Valley hosts the country’s largest concentration of banks, government agencies, telecommunications providers and multinational corporations.

Certain facilities supporting financial trading, enterprise cloud services, disaster recovery and network interconnection benefit from being close to customers and major fibre networks.

RHB Research says urban areas require DCs designed with high-tier resilience to support production and mission-critical workloads. “Such facilities need to be located close to enterprise headquarters, regulators and key networks.”

Lum Chean Jeeng, director of commercial and industrial solutions at AQ Energy, tells StarBiz 7 that urban locations may attract greater DC interest, partly because they are typically more infrastructure-ready, especially in terms of power supply.

“DCs have to meet strict uptime of 99.99%, so proximity to power supply is extremely crucial.

“You can still build hyperscale facilities in urban areas, provided you upgrade infrastructure such as cables connecting substation to DCs. Tenaga Nasional Bhd would typically pass this cost on to DC developers,” says Lum.

Urban DCs can strengthen Malaysia’s digital infrastructure resilience while supporting businesses that require reliable, low-latency connectivity. Yet proximity should not become a blanket justification.

Many hyperscale cloud facilities process workloads that are neither customer-facing nor particularly sensitive to latency.

AI training centres, for example, prioritise access to large amounts of electricity far more than proximity to office towers or residential neighbourhoods.

Gary Goh, founder and director at Sprint DC Consulting, concurs that not all DCs need to be located close to urban areas.

“But the reality is that the DC market is very tight now in Malaysia. Interested tenants are finding it difficult to secure available DC capacity.

“So, some of them will have to take space in urban DCs as long as there are vacancy, despite the fact that they don’t have to be located so close to the city.”

For now, urban DCs are not restricted in terms of the types of tenants they may serve.

Now, this raises a broader question: Did the urban planners consider this before approving such projects?

At the policy level, the Data Centre Task Force (DCTF) provides strategic direction. The inter-agency body is co-chaired by the Investment, Trade and Industry Ministry (Miti) and the Digital Ministry.

Deputy Miti Minister Sim Tze Tzin told Parliament that the DCTF comprehensively studies the entire DC ecosystem, examining each application based on factors such as power and water supply capacity before approval is granted.

He assured that the government will only approve DC projects after ensuring that energy and water supply capacity is sufficient to meet the needs of the people and local industries.

Earlier in February this year, Prime Minister Datuk Seri Anwar Ibrahim announced that the government has restricted approvals for new non-AI DCs, citing rising energy and water consumption.

Apart from clearance from the DCTF, DC projects are also subject to a range of existing laws covering land, planning and construction, utilities, and operational licensing, including requirements under the Communications and Multimedia Act 1998.

Regardless, Treasury secretary general Tan Sri Johan Mahmood Merican recently acknowledged that Malaysia’s DC boom has exposed policy gaps.

“I think in recent experience, we perhaps overdid the red carpet for DCs to the extent that now we are almost getting too much of a demand,” he was reported as saying.

Too close to neighbours?

While Malaysia has put regulations in place, it is worth asking whether these frameworks fully account for DCs and their potential impacts.

One concern is the requirement for a 50m buffer zone, comprising features such as roads, building setbacks, perimeter planting or parking, between DC structures and residential lot boundaries.

The 50m buffer zone is part of PlanMalaysia’s Data Centre Planning Guidelines 2024.

Greenpeace Malaysia’s campaigner Yan says the 50m buffer reflects an “oversight” of the potential negative impacts of allowing DCs to be built in close vicinity to homes.

The policy concern is not merely whether 50m is sufficient, but whether a single buffer distance can adequately regulate DCs ranging from small facilities to hyperscale complexes.

Ihsan Zainal Mokhtar, a registered town planner at IZM Consult, tells StarBiz 7 that it is important for DC locations to be properly determined.

He believes there is still a lack of spatial planning analysis on locations of DCs in urban areas. Spatial planning is the process of deciding where different types of development should be located and how land should be used.

“This is important because we still have basic issues like public facilities and amenities, particularly in urban centres.

“Spatial planning is the manifestation of the various aspects of town planning, including economic, social and political considerations,” he adds.

Greenpeace Malaysia, in a response to StarBiz 7, echoes calls for DCs not to be built in residential areas. Its community associate coordinator Aidil Iman says the development of DCs could risk the loss of green spaces.

“When operating, DCs could affect the long-term health of community members through noise pollution and heat generation,” he says.

Aidil also points to a report by Greenpeace Australia Pacific, which found that DCs have continued to rely on non-renewable energy (non-RE) sources for power generation, including gas, which further derails Australia’s energy transition plans.

On the surge in power demand, Greenpeace Malaysia’s Yan says policymakers appear to believe that a two-pronged strategy will be sufficient – optimising PUE and related metrics, alongside a shift toward RE.

“Much remains guidance rather than enforceable regulation. There is no specificity on defined targets for emission reductions expected from energy-efficient hardware, AI-optimised cooling or ‘green’ DC design and operations,” she points out.

She adds that without binding targets and consistent enforcement, the country’s largely voluntary and fragmented policy approach leaves a significant gap that needs to be addressed.

“In turn, this raises the question of whether the current guidelines can genuinely be considered ‘sustainable’, given the risks of greenwashing and the omission of public safety and interest considerations,” Yan says.

Goh points out that capital expenditure (capex) will be higher for DCs seeking to achieve lower noise levels.

“DCs in urban locations sometimes have to double down on noise reduction beyond the regulatory prescribed level to avoid conflict with residential communities.”

Gains must be sustainable

Costs of compliance aside, another key question is opportunity cost: What cities forgo when scarce urban land is allocated to DCs?

A parcel in Petaling Jaya or Cheras could instead support offices, mixed commercial developments, healthcare facilities or higher-density housing.

Such developments typically generate greater employment, stronger surrounding economic activity and more vibrant urban environments.

DCs produce a different economic profile. Construction creates significant business opportunities for contractors, engineers and equipment suppliers.

Once completed, however, operations are highly automated, and permanent staffing levels are relatively modest compared with the scale of capital invested.

A report by Amro highlighted that the lion’s share of jobs created by DCs takes place during the pre-operational stages. Once operations begin, most DCs create only 30 to 50 permanent jobs on average, it said.

Investment announcements, therefore, tell only part of the story.

Malaysia has understandably celebrated billions of ringgit in approved DC investments. However, less attention has been focused on what should be more pertinent questions:

How many long-term Malaysian jobs will each facility create? How much local procurement will remain after construction? How much tax revenue will be generated?

How many Malaysian companies will actually benefit from the computing capacity?

Take Johor, the leader in DC investments. In 2025, despite being the only state to enjoy approved investments of above the RM100bil mark, Johor was expected to generate some 24,584 potential jobs from these investments.

Compare this with Penang, the country’s manufacturing leader. Its approved investments of RM32.9bil in 2025 were expected to create 28,570 jobs – outpacing Johor.

Put simply, every RM1mil of approved investment in Penang could create 0.87 jobs, while every RM1mil of approved investment in Johor could create only 0.22 jobs.

Any expert would agree that there is a risk of equating capex with lasting economic value.

Safeguarding resources and communities

Resource allocation presents another challenge. DCs consume large amounts of electricity and require exceptionally reliable power supply.

Every additional megawatt committed to a DC represents capacity that cannot simultaneously serve other industrial users without further investment in generation and transmission.

Electricity demand itself is not a problem if supply expands accordingly.

The issue is whether public infrastructure is being upgraded efficiently and whether limited grid capacity is being directed towards activities that generate the greatest economic return.

Water deserves similar attention. Modern cooling technologies have reduced consumption, but efficiency ratios alone do not answer the broader policy question.

Urban communities will naturally ask whether industrial cooling should compete with domestic needs during periods of supply stress.

S&P Global Ratings says that Malaysia’s power and water reserve margins are still above healthy levels. It also expects power generation capacity additions to keep pace with rising demand from DCs.

However, this does not mean DC approvals do not require scrutiny and policy adjustments.

Neighbourhood concerns should also be viewed through an economic rather than emotional lens.

Noise from cooling systems, backup generators, construction traffic and visual impact all represent external costs borne by surrounding communities.

Such costs rarely appear in investment announcements, yet they form part of the project’s overall economic equation.

None of this argues against DCs. Malaysia requires more digital infrastructure if it intends to compete in AI, cloud computing and the wider digital economy.

Location, however, should become a far bigger part of the national conversation. Malaysia would benefit from a more sophisticated approval framework.

Every urban DC should be required to demonstrate commercial necessity, quantify its long-term economic contribution and disclose its impact on electricity, water, transport and neighbouring communities before approval is granted.

Digital infrastructure is essential for economic growth. Choosing the right location for that infrastructure is equally important.

Malaysia’s objective should not simply be to build more DCs.

It should be to ensure every urban DC justifies the scarce land, energy and public infrastructure it consumes, while delivering measurable economic value that extends well beyond the construction phase.