Pranav Hotkar
Beijing, China - March 17, 2026 - Google is in talks with multiple suppliers, including Envicool, to secure liquid cooling systems for its expanding data center footprint, as surging AI demand strains global infrastructure supply chains.
According to sources familiar with the matter, Google’s procurement team recently visited China to meet with cooling equipment manufacturers, signaling a push to diversify suppliers amid tightening availability. The discussions focus on advanced liquid cooling technologies, which are increasingly critical for supporting high-density AI workloads.
Unlike traditional air-based systems, liquid cooling uses circulating fluids to dissipate heat more efficiently, making it essential for modern AI servers powered by high-performance GPUs and custom accelerators. As compute density rises, hyperscalers are rapidly shifting toward these systems to maintain performance and energy efficiency.
The move highlights a growing supply bottleneck beyond semiconductors, with cooling infrastructure emerging as a key constraint in scaling AI data centers. Industry estimates suggest the global liquid cooling market is set to nearly double in value in 2026, driven largely by demand from hyperscale AI deployments.
Envicool, one of the suppliers in discussions, manufactures coolant distribution units (CDUs), a critical component that delivers liquid coolant directly to server racks. Some of these systems are being tailored to meet Google’s specific design requirements, underscoring the customized nature of next-generation data center infrastructure.
Google’s outreach to Chinese vendors also reflects broader supply chain dynamics. With limited production capacity in traditional manufacturing hubs such as Taiwan, companies are increasingly turning to alternative regions to secure components at scale. Chinese suppliers, supported by strong domestic demand and manufacturing capabilities, are becoming key players in the global cooling ecosystem.
However, the development carries geopolitical implications, as US technology firms deepen engagement with Chinese hardware providers amid ongoing trade and technology tensions. This adds complexity to procurement strategies even as demand pressures mount.
The urgency behind these talks stems from the rapid expansion of AI infrastructure, where cooling has become as critical as compute and power. Without sufficient thermal management, high-density AI clusters cannot operate reliably, making liquid cooling a foundational requirement rather than an optional upgrade.
As hyperscalers race to deploy next-generation AI capacity, Google’s efforts to secure cooling systems highlight a broader industry shift: the AI data center stack is no longer defined by chips alone, but by the availability of power, cooling, and integrated infrastructure at scale.
