Shell Unveils Certified Fluid for Future Data Centers

Shell Unveils Certified Fluid for Future Data Centers

18 Oct 2025

Rapid adoption of artificial intelligence is straining data centres across the Middle East, where high temperatures make cooling both costly and complex. On May 13th Intel certified Shell’s gas-to-liquid (GTL) single-phase immersion cooling fluids for use with its 4th and 5th generation Xeon chips. It is the first such approval for this type of fluid and gives operators more confidence to use immersion systems at scale.

Shell pressed its case in Dubai at GITEX GLOBAL, held on October 13th to 17th. Executives argued that immersion cooling can trim energy use while supporting denser racks, an attractive pitch in Gulf states where electricity demand from data centres is rising fast. The region’s push to host AI models makes the search for efficient cooling less a choice than a requirement.

The firm says that immersion systems can cut electricity use by as much as 48% compared with air cooling, though savings depend on site design and workloads. Lower consumption means lower costs and, in theory, lower carbon emissions. It also means operators can squeeze more computing into each square metre, delaying the need for new buildings.

Shell is keen to avoid confusion between its products. Its immersion fluids, such as the S-series, are dielectric liquids designed to submerge servers. They are distinct from its Diala transformer oils, which are made for high-voltage equipment. The certified fluids are already used in various enterprise settings outside the Middle East, giving the technology a modest track record.

For the region, the implications are simple enough. As governments and cloud firms race to build capacity for AI, immersion cooling offers a practical route to higher densities without relying on ever larger air conditioning loads. Intel’s endorsement lowers perceived risk for early adopters.

Yet progress will not be effortless. Retrofitting older sites, training staff and managing fluids will all require capital and discipline. Even so, the mix of technical validation, rising thermal stress and intensifying AI ambitions suggests that more pilots, and perhaps wider deployment, are likely in the year ahead.