Frore Systems today launched LiquidJet Nexus, a groundbreaking liquid cooling solution tailored for AI servers like NVIDIA Kyber. As AI compute densities surge, traditional cooling methods struggle to manage heat, creating a critical bottleneck. LiquidJet Nexus addresses this by integrating multi-stage 3D short-loop jetchannel coldplates into a unified system, delivering 75% higher heat transfer efficiency and reducing thermal stack weight by 65%. Components stay 8°C cooler, eliminating complex hoses and manifolds while enabling 53°C inlet temperatures—a milestone that cuts reliance on mechanical chillers and slashes data center power use by 10%.

The system redefines the AI Thermal Stack, a critical infrastructure layer for next-gen servers. Traditional liquid cooling relies on cumbersome assemblies of hoses and connectors, but LiquidJet Nexus simplifies this with a single, lightweight design. This reduces assembly complexity, improves reliability, and supports 2x higher compute density per rack. Hyperscale data centers can now scale AI workloads without thermal limitations, lowering total cost of ownership.

Cooling has evolved from an afterthought to a foundational AI infrastructure element. As AI models demand exponential compute growth, thermal constraints threaten to stall progress. Frore’s LiquidJet Nexus tackles this by enabling higher power densities and energy efficiency. The company also extends its thermal expertise to edge devices with AirJet, a solid-state active air-cooling chip for compact, silent, and dustproof systems. Together, these technologies bridge the gap between cloud-scale AI and edge deployments.

"Cooling is no longer just an afterthought," said Dr. Seshu Madhavapeddy, Frore Systems’ founder. "It is about enabling the next generation of AI infrastructure." With LiquidJet Nexus, the industry gains a scalable solution to power AI’s thermal demands, ensuring performance and efficiency align with the software and compute stacks driving the AI revolution.