Wiwynn Shows Vera Rubin, AMD Helios Racks and Diamond Cooling at OCP 2026

Rack platforms for NVIDIA and AMD

Wiwynn is presenting at Booth C74 of the San Jose Convention Center. According to the company, it and Wistron are among the first NVIDIA Pilot Build Partners for the NVIDIA Vera Rubin NVL72 platform, which is used for engineering validation and design-in. The liquid-cooled, rack-scale architecture integrates 72 NVIDIA Rubin GPUs and 36 NVIDIA Vera CPUs with 6th-generation NVIDIA NVLink.

The company also shows the AMD Helios rackscale solution, built on the Open Rack Wide (ORW) standard. Helios combines AMD Instinct MI450 Series GPUs, 6th Gen AMD EPYC Server CPUs, AMD Pensando networking technologies and AMD ROCm open software for large-scale AI training and inference. Wiwynn says it applies its rack integration and manufacturing expertise to bring the architecture to customers at scale, with solutions tailored to cloud service provider requirements.

"As AI workloads push compute density to unprecedented levels, the industry needs partners who can deliver proven, full-stack engineering across the entire rack," said William Lin, President and CEO of Wiwynn.

Cooling: from diamond to lighter cold plates

A large part of the showcase is thermal engineering. Wiwynn lists several designs:

• Diamond-based microchannel cooling: the company is exploring diamond-based composite materials as an alternative to copper, demonstrating direct diamond-to-silicon bonding with integrated microchannel structures for on-package cooling. It says diamond's thermal conductivity shortens the heat transfer path at the chip level.
• 6kW double-sided cold plate: an ultra-thin design with cavity PCBA for high-power AI ASICs using Vertical Power Delivery. Wiwynn claims 6kW cooling capacity, over 15% thermal performance improvement and up to 80% reduction in power delivery loss.
• Lightweight cold plate and inner manifold: by using less metal and lightweight tubing, Wiwynn says it cuts cooling component weight by at least 30%, which eases integration, transportation and installation.
• Topology-optimized microchannel cold plate: developed with Corintis, these designs are tailored to chip power maps and target high heat flux regions while balancing thermal and hydraulic performance.

The company says the topology-optimized approach has the potential to further improve cold plate performance for next-generation AI ASICs.

Co-packaged optics for network bottlenecks

Together with partners including Ayar Labs, Corning, GUC and Senko, Wiwynn shows a co-packaged optics (CPO) scale-up rack solution. It covers rack-level fiber routing, patch panel integration and multi-vendor fiber management, and includes a live demonstration of optical engine telemetry.

According to Wiwynn, the demo highlights an open ecosystem approach and its ability to integrate optical, mechanical, cooling and system-level innovations into a scalable rack-scale AI infrastructure platform.

What it means for customers

The announcement positions Wiwynn as a supplier offering design through high-volume rack integration for cloud service providers. The company describes its capabilities as spanning server design, system integration and high-volume L10/L11 rack delivery, backed by a manufacturing network in Taiwan, the United States, Mexico, Malaysia and the Czech Republic.

Visitors can follow the technologies in Executive Track and Expo Hall Talk sessions, engineering workshops and the Future Technology Symposium. Wiwynn also offers whitepapers on AI infrastructure and data center technologies. The source gives no pricing or availability details for the showcased products.