Palm Oil and AI: An Odd Pairing I’m Watching, Not Buying
I came across something last week that doesn’t fit anywhere in my portfolio, which is partly why I’m writing about it today.
The Malaysian Palm Oil Board (“MPOB”) has been promoting a technology called Sawit EcoTherm — a palm-based, non-conductive coolant for high-performance infrastructure, including data centres. The claim is that it can improve cooling efficiency while cutting down on the water used by conventional cooling systems.
Palm oil and AI infrastructure are two industries which I would not normally put together. But AI chips are pushing up power density and heat output. Immersion cooling, putting AI equipment into a non-conductive fluid instead of water-based systems, is one possible approach being explored to deal with those issues.
The water part is what made me look twice.
Malaysia’s water problem is real
Malaysia has become Southeast Asia’s biggest data-centre construction market, most of it concentrated in Johor. The Energy Transition and Water Transformation Ministry projects data centres across Selangor, Johor and Negeri Sembilan could need 445 million litres of water a day by 2030, up from about 29 million litres a day as of January 2026.
That is a lot of water.
It has already gotten politically sensitive. Residents in Gelang Patah, Johor, had protested near a construction site earlier this year over dust and water concerns, even as the Johor government separately had to clarify that some unrelated water disruptions were caused by drought, not data centres.
The industry is already responding. Several Johor operators have shifted to reclaimed and treated wastewater instead of drawing clean water from the public supply. So Sawit EcoTherm isn’t solving a problem nobody’s thought about. It’s another option pointed in the same direction.
Is the science real?
I did a little research, and this is what I discovered.
Palm oil as a dielectric coolant is not actually a new idea. There is academic research going back well over a decade comparing palm oil to mineral oil on thermal and insulating properties, mostly out of the transformer-insulation field rather than data centres. One study found crude palm oil absorbed heat slightly more effectively than mineral oil under certain conditions.
So the underlying science isn’t coming out of nowhere.
However, I couldn’t find an independent or peer-reviewed study on Sawit EcoTherm itself — only MPOB’s own statements. MPOB says the technology is currently being offered for commercialisation and is IP-protected.
Who actually benefits
If this becomes a real product category, I do not think the obvious winners are the plantation companies.
The more interesting question is whether AI creates demand for a higher-value palm derivative, not just another tonne of crude palm oil. That is where downstream processors could capture more of the economics than a plantation-only business.
Wilmar (SGX: F34) came first into my mind — plantations, refining, oleochemicals, and has a large integrated platform. IOI Corporation (Bursa: 1961, IOICORP) is smaller but structured similarly. ADM (NYSE: ADM) has palm exposure too, though it’s a small slice of a much bigger agricultural business.
Of the three stocks, IOI is the one that sticks in my head.
Not because I think it is about to become an AI infrastructure company. It is not. IOI would be several steps removed from any of the AI infrastructure names I actually own — at most, a very different kind of AI-adjacent commodity bet.
And here’s why — the Malaysia pattern
Worth mentioning: Malaysia has shown with BYD that it’s willing to shape the terms under which foreign companies operate in strategically important industries.
BYD (1211.HK) announced its Tanjong Malim CKD plant in August 2025, fully self-funded, production targeted for late 2026. A month later, the Ministry of Investment, Trade and Industry (“MITI”) approved the manufacturing licence — but attached conditions. A domestic sales cap. A minimum price on locally sold units. MITI’s own statement says the price floor exists to preserve market space for Proton and Perodua.
As of a couple of weeks ago, MITI told parliament it still had not heard from BYD on whether the plant is proceeding. BYD executives have been spotted looking at a contract-assembly plant instead.
BYD was never blocked. It was let in, and the terms around its participation subsequently changed in a way that benefited the incumbents.
I do not know if anything similar would happen with Sawit EcoTherm. MPOB has not said anything about conditions or partners. But the BYD pattern is why I would not assume that the biggest global balance sheet automatically wins this one. A foreign partner like Wilmar or ADM is not excluded by default — but local sourcing or licensing conditions could still end up favouring a Malaysian processor, IOI included, whether or not IOI is the named partner.
And even without a named-partner outcome, Wilmar’s scale means it probably captures some of the demand anyway, if palm-derived coolants become a real category at all.
I’m not buying it at this moment
No confirmed partner. No deployment. No way to size the economics yet.
It currently does not sit inside any of the four themes my portfolio actually runs on.
It is likely going onto my 2027 general watchlist. When I sit down to plan next year, I will check whether there’s anything real to size. If not, it drops off.
Disclaimer: This post is for informational purposes only and does not constitute financial advice. I am not a licensed financial advisor. Please do your own research and consult a qualified professional before making any investment decisions.