Wiwynn Unveils Next-Gen AI Platforms and Cooling Technologies

Wiwynn Corporation presented its latest AI infrastructure innovations at the 2026 OCP Global Summit in San Jose, focusing on next-generation AI rack platforms, advanced thermal engineering, and optical interconnect technologies. The company highlighted its comprehensive approach across compute, cooling, power, and connectivity.

The showcased AI rack platforms included the NVIDIA Vera Rubin NVL72, a liquid-cooled, rack-scale architecture integrating 72 NVIDIA Rubin GPUs and 36 NVIDIA Vera CPUs. Wiwynn also presented the AMD Helios solution, built on the Open Rack Wide standard, combining AMD Instinct MI450 GPUs, AMD EPYC CPUs, and AMD ROCm software for large-scale AI training and inference.

Advancements in thermal engineering were also a key focus, with Wiwynn demonstrating diamond-based microchannel cooling for on-package chip applications and a 6kW double-sided cold plate designed for high-power AI ASICs. The company also presented lightweight cooling components and topology-optimized microchannel cold plate designs.

Additionally, Wiwynn, in collaboration with partners, showcased co-packaged optics (CPO) based solutions for rack-level optical interconnects. These solutions aim to address network bottlenecks by integrating optical, mechanical, and cooling innovations into scalable AI infrastructure platforms.