LG Uplus and Ecolab Korea are joining forces to develop operating standards for liquid cooling in AI data centres, with a test environment and plans for AI-based predictive maintenance. The partnership targets a practical pressure point in AI infrastructure: keeping increasingly powerful systems cool without letting coolant or equipment problems turn into hardware failures.
Watch Desk analysis
What happened
The companies signed a memorandum of understanding in Seoul. LG Uplus will provide a large-scale liquid-cooling test environment at its Pyeongchon 2 data centre, while Ecolab Korea will develop procedures to monitor coolant quality. They also plan to analyse operating data, set standards and detect equipment problems early.
The Korea Times reports that the companies intend to develop AI-based predictive maintenance systems to identify irregularities before they lead to failures. Read the report.
Why it matters
AI data centres pack substantial computing power into limited space, producing heat that conventional air cooling may struggle to manage efficiently. Liquid cooling can help, but it brings its own maintenance demands: deteriorating coolant or poorly maintained equipment can undermine cooling and damage hardware.
This partnership puts attention on the operating standards and monitoring that make cooling dependable, not just on the headline hardware. If the planned test work produces useful procedures, it could help operators spot trouble earlier. The predictive-maintenance system is still a plan, not a demonstrated result.
Our read
AI infrastructure is often discussed in terms of chips and power. Cooling is less glamorous, but a GPU that overheats is not much of an infrastructure strategy. The useful part here is the combination of a real test environment, coolant monitoring and operating-data analysis. We’ll be looking for evidence that this work produces standards operators can actually use, rather than another promising acronym cooling its heels in a slide deck.
What to watch
- What systems the companies test at the Pyeongchon 2 data centre.
- Whether they publish practical operating standards for liquid cooling and coolant quality.
- Whether the proposed predictive-maintenance system is built and shown to detect faults early.
Discussion spark: Should AI data-centre operators treat liquid-cooling standards as shared industry infrastructure, or let each provider develop its own approach?
Sources and evidence
- Source update (28 October 2026, 02:36 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.