ZutaCore Ltd., the Israeli developer of two-phase, waterless liquid cooling for high-density servers, has raised $100 million in a Series C funding round to scale deployments as AI chips push data-center thermals beyond what air and water systems were ever designed to handle.
Strategic backers from cooling, chips and motors
The round drew strategic capital from Mitsubishi Electric Corp., Carrier Ventures and Samsung Venture Investment Corp., the corporate venture arm of Samsung Electronics, alongside additional investors. Goldman Sachs & Co. acted as exclusive placement agent. The deal lands as hyperscalers commit tens of billions to new AI capacity and operators scramble to retrofit existing halls.
HyperCool: dielectric fluid that boils at the chip
ZutaCore's HyperCool platform is a sealed, closed-loop system that pumps a nonconductive dielectric fluid through cold plates mounted directly on processors. The fluid boils on contact with the silicon, carrying heat away as vapor to a coolant distribution unit where it condenses and recirculates. Because the fluid is nonconductive and enclosed, a leak does not short electronics, eliminating a long-standing barrier to liquid cooling in production fleets.
Built for 4,000-watt-plus AI silicon
The system is engineered to cool processors drawing more than 4,000 watts, well beyond air cooling's limits and at the edge of single-phase liquid systems. Cold plates are compatible with Intel, AMD and Nvidia chips, including a new OmniTherm plate that brings two-phase cooling to Nvidia's RTX PRO 6000 Blackwell Server Edition in a single-slot PCIe footprint.
Skipping the water bill
The pitch leans on what HyperCool removes. Traditional evaporative cooling can consume millions of gallons of water per year, attracting regulatory pressure and local opposition in arid regions. ZutaCore says its closed-loop architecture cuts cooling energy use roughly in half versus conventional approaches and can be retrofitted into existing racks — a contrast with the bespoke builds typical of today's most advanced AI factories.
75 deployments, megawatt scale on tap
The company says it now has more than 75 deployments across the Americas, Europe and Asia and has built a 2-megawatt end-of-row emulation platform in Israel to validate large-scale customer rollouts. Chairman and CEO Erez Freibach said the new capital will fund global commercialisation, additional R&D and support for megawatt-class installations.
Reporting based on coverage from SiliconANGLE (Duncan Riley, 2 June 2026), HPCwire and Globes.