Malaysian Original · Science, palm and digital infrastructure

Sawit EcoTherm

Bangi, Selangor

Inside the Malaysian research effort to turn a palm-derived fluid into cooling for AI servers—and the difficult proof still needed outside the laboratory.

AI-generated editorial interpretation of computer-server boards immersed in golden dielectric fluid inside a tropical research laboratory — AI-generated editorial visual · not documentary photography

Evidence status · Research-stage technology; public pilot data not yet available

2026Introduced for transfer

01 · The opening scene

A server disappears into golden liquid.

The first instinct is to pull it out. Electronics and liquid are trained enemies; a spill near a laptop is an emergency. Yet in a laboratory in Bangi, the liquid is the point. A computer board can sit beneath the surface while heat leaves its most demanding components and moves into the surrounding fluid.

The fluid is called Sawit EcoTherm. Developed by researchers at the Malaysian Palm Oil Board, it uses palm-derived material to perform a job usually associated with petroleum-based dielectric liquids: cool high-performance computing hardware without conducting electricity.

That pairing makes the idea instantly memorable—Malaysian palm oil meeting the physical machinery of artificial intelligence. It also makes the story easy to exaggerate. Sawit EcoTherm has technical specifications and a government research programme behind it, but no named live data-centre pilot with public operating results. The real story sits between those facts: invention has happened; proof at infrastructure scale has not.

AI-generated editorial interpretation of a palm-derived cooling-fluid research workbench with glassware and an unbranded server board — AI-generated editorial visual · not documentary photography

Developed by: Malaysian Palm Oil Board researchers
Introduced: MPOB technology-transfer programme, June 2026
Purpose: Single-phase immersion cooling
Current stage: Seeking pilots, validation and commercial partners

The heat behind the cloud

AI has a physical temperature.

The word cloud hides racks, cables, pumps and heat. Every processor doing the dense calculations behind an AI service uses electricity, and much of that energy eventually becomes heat that must be moved away. As chips become more powerful and tightly packed, conventional air cooling faces a harder task.

Many data centres use chilled water or evaporative systems somewhere in that chain. The exact water impact depends on climate, design, power generation and accounting boundaries, so no single statistic describes every facility. What is clear is that cooling choices now belong in discussions about where digital infrastructure is built and what resources it consumes.

Immersion cooling offers a different geometry. Instead of blowing cold air across server components, hardware is placed in a non-conductive fluid. Heat passes directly into the liquid and can be carried around a closed loop towards a heat exchanger. The concept is established; Sawit EcoTherm's novelty is the proposed palm-based fluid inside that system.

The cloud does not float. It occupies land, draws power and needs somewhere for its heat to go.

A Malaysian material meets a new market

Palm moved from commodity to engineering input.

Malaysia's palm story is usually told through cooking oil, exports and arguments about land use. MPOB's technology-transfer programme asks a different question: can palm-derived oleochemicals become higher-value industrial materials? In June 2026, the board introduced Sawit EcoTherm among nine technologies offered for commercialisation.

The public technical sheet names three researchers—Noor Khairin Mohd, Zafarizal Aldrin Azizul Hasan and Nik Nurfatmah PZ Nik Pauzi—and describes two clear formulations for single-phase immersion cooling. It reports dielectric breakdown voltage above 75 kilovolts as well as low water content and acidity. Those measurements help establish that the formulations exist and have relevant laboratory characteristics.

They do not, by themselves, establish a sustainable data centre. The fluid would operate as one component in a larger system of tanks, pumps, heat exchangers, servers and power supply. Its environmental result would also depend on feedstock origin, manufacturing energy, service life, maintenance and what happens at end of use.

A renewable feedstock is the beginning of an environmental case, not the conclusion.

The researchers' wager

The most interesting claim is not ‘palm oil cools a server.’

Palm-derived liquids can carry heat; the technical sheet supports a plausible route. The more ambitious wager is that Malaysia can combine its oleochemical knowledge with its growing digital-infrastructure sector and create more value locally than it would by exporting another undifferentiated commodity.

That makes Bangi important. MPOB is not merely attaching a national ingredient to a fashionable industry. Its researchers are trying to translate decades of palm science into a specification that data-centre engineers can test. If the work succeeds, the intellectual property, technical expertise and manufacturing knowledge could matter as much as the feedstock.

The research also forces two Malaysian development stories into the same room. Data centres promise investment and digital capacity but raise questions about power and water. Palm supports livelihoods and industrial capability but carries a global environmental burden. A credible solution must answer the difficult parts of both stories rather than using one to green the image of the other.

The national opportunity is not a golden liquid. It is the engineering knowledge built around it.

From a specification to a system

Now comes the trial that can prove—or shrink—the promise.

MPOB has introduced Sawit EcoTherm for technology transfer, not announced a finished consumer product. In a Bernama interview, the research team identified pilots, technical validation and evidence of long-term performance, reliability and cost as the next priorities. That candour is essential.

A useful pilot would name the facility, workload and system size; compare energy and water use against a clear baseline; track fluid condition and hardware compatibility over time; and disclose enough methods for specialists to interrogate the result. Fire behaviour, ecotoxicity, biodegradation and end-of-life handling also need independent scrutiny.

The most persuasive outcome may be narrower than the headline. Sawit EcoTherm does not need to cool every AI server or eliminate all water use. It needs to demonstrate where it performs safely, what resources the whole system consumes and whether its lifecycle impact improves on a credible alternative. That is how a striking Malaysian idea becomes infrastructure rather than content.

A good pilot does not decorate the promise. It gives the promise somewhere it can fail.

06 · Why this story matters now

Celebrate the question. Keep testing the answer.

Sawit EcoTherm deserves attention because it is a locally specific response to a global engineering problem. It brings Malaysian researchers, an established national material base and the infrastructure behind AI into one ambitious experiment.

It also models the kind of Malaysian Original this site should follow over time, not merely announce once. If pilots succeed, the story becomes one of commercialisation and measurable resource savings. If they expose limits, those results still deepen local technical knowledge.

For now, the right posture is informed optimism. The formulation is real. The laboratory specifications are worth examining. The environmental and operating claims remain propositions to test. Shouting about Malaysian ingenuity is most credible when curiosity and scrutiny rise together.

07 · The evidence check

What the public record shows.

Sawit EcoTherm is a documented MPOB research technology with published formulation specifications. The strongest performance and sustainability claims still come from its developers rather than an independent operating trial.

Documented progress

01
MPOB introduced Sawit EcoTherm in June 2026 among technologies offered for commercialisation.
02
Technology Transfer sheet TT702 identifies two clear palm-based dielectric-fluid formulations intended for single-phase immersion cooling.
03
The sheet reports breakdown voltage above 75 kV along with low water-content and acidity specifications.
04
MPOB's researcher says the next priorities are pilot projects, technical validation and proof of long-term performance, reliability and cost.

The commercialisation test

01
A named live data-centre pilot with its scale, workload, duration, operating baseline and measured results.
02
Independent evidence for long-term server compatibility, fire safety, biodegradation, ecotoxicity, fluid service life and end-of-life handling.
03
A full lifecycle assessment—including traceable palm feedstock—and a clear definition of energy and water savings at whole-facility level.

Editorial position

This is an emerging technology, not a visitor attraction or product recommendation. ‘Water-free cooling’ should be understood as a potential facility design enabled by immersion cooling—not a demonstrated property of the fluid alone. We will update this profile when a named pilot and independently reviewable operating data become public.

Read the MPOB announcement ↗