IT Brief UK - Technology news for CIOs & IT decision-makers
Flux result 288f3bce a5d7 4ce4 8c25 4a5733cb54b1

Iceotope passes 200 patents as AI cooling demand rises

Thu, 16th Apr 2026

Iceotope has passed 200 granted and pending patents in liquid cooling, with 109 granted patents and 99 pending applications.

The milestone comes as operators of data centres and edge computing sites face rising heat loads from artificial intelligence systems and other dense computing equipment. Iceotope estimates the global market for data centre cooling will reach $40 billion to $45 billion by 2030, with liquid cooling accounting for $15 billion to $20 billion.

Founded in 2005, the company has focused on liquid cooling systems for servers and other computing hardware. Its patent portfolio covers chassis design, the use of dielectric fluid, and rack-scale thermal management.

Iceotope's approach uses a sealed server or component and a small amount of non-conductive dielectric fluid to cool heat-producing parts inside the system. The company describes the method as a form of precision liquid cooling that combines features of direct-to-chip and immersion cooling.

According to Iceotope, the design works with standard racks and existing data centre infrastructure. The systems can also be deployed for artificial intelligence and high-performance computing workloads at the edge, including in harsher operating environments.

Cooling Pressure

Thermal management has become a bigger issue as chips grow more powerful and rack density rises. Air cooling remains common across much of the sector, but newer artificial intelligence deployments have increased interest in alternatives that can remove more heat while limiting energy and water use.

Iceotope says its systems can reduce energy use by up to 40% and water use by up to 96% compared with traditional cooling methods. The design also allows higher rack density, an increasingly important factor as operators try to make better use of expensive data centre space.

The company says its intellectual property was developed through work with hyperscalers, silicon providers, and original equipment manufacturers. It is intended to address practical deployment issues rather than remain limited to laboratory designs.

Iceotope also points to its focus on cooling all heat-producing elements within a server rather than concentrating mainly on processors. That includes processors, power supply units, storage, and networking components.

This differs from cold-plate approaches, which are generally aimed at flat chip surfaces and key hotspots. Iceotope argues that a broader cooling method can support environmental resilience and quieter operation while managing heat across the full compute stack.

Early Start

Iceotope says it began developing liquid cooling systems well before artificial intelligence infrastructure became a major focus for corporate boards and investors. The company anticipated that heat generated by denser computing equipment would eventually exceed the practical limits of conventional air cooling.

Its current patent count points to a long-running effort to secure protection across multiple parts of the design and deployment chain. In sectors such as data centre infrastructure, patent portfolios can influence licensing, manufacturing partnerships, and product adoption, particularly where equipment must be integrated with existing server and rack designs.

Iceotope also says it has worked with major artificial intelligence infrastructure original equipment manufacturers, hyperscalers, semiconductor providers, and supercomputing centres. It did not name those organisations in the announcement.

Neil Edmunds, chief innovation officer at Iceotope, said the patent tally reflected years of engineering work on practical cooling problems.

"This milestone reflects the depth of innovation and expertise of our team," Edmunds said. "Each patent represents a real engineering challenge solved, whether that is cooling denser chips and more powerful components, integrating into existing infrastructure, or reducing the environmental footprint of AI workloads. As demand for data centre and edge cooling continues to grow, this IP is the foundation that allows us to stay ahead of the industry."