Google reportedly sent Taiwan procurement team to mainland to buy AI data‑centre liquid‑cooling gear
What happened
It has been reported that a procurement team from Google’s Taiwan unit visited mainland China this month to seek suppliers of data‑centre liquid‑cooling equipment, meeting with Shenzhen Yingweike (英维克科技) and planning talks with at least one other Chinese firm, according to Reuters. The visit, said sources, reflects growing strain in the supply chain for liquid‑cool components as cloud providers race to build out high‑density AI infrastructure.
Why it matters
Liquid cooling — which circulates water or other fluids around or above servers — has become essential as AI racks cram more compute into smaller footprints and traditional air‑cooling can’t cope. JPMorgan estimates the global market for AI server liquid‑cooling systems will surge from about $8.9 billion last year to more than $17 billion by 2026. So who makes the pumps, manifolds and racks matters almost as much as who makes the chips.
Geopolitics and supply risk
This procurement push underscores a second, less visible front in the AI race: lower‑value, high‑volume hardware is also in short supply. Reportedly, the visits show Chinese suppliers are increasingly integral to global data‑centre expansion. That matters against a backdrop of U.S. export controls and wider tech decoupling — restrictions that target advanced semiconductors but leave room for equipment flows in other parts of the stack. Can Western cloud firms secure enough non‑chip hardware without new sourcing strategies? The answer will shape how fast global AI deployments scale.
