Verda and Compal Expand AI Hardware Manufacturing
The partnership increases production capacity for high-density NVIDIA Blackwell Ultra server platforms globally.
Updated on Oct. 21, 2026 in Data Centers
Live Poll
Do you believe domestic manufacturing expansion by large technology companies is good for the country?
Verda and Compal are expanding their AI infrastructure partnership to increase global manufacturing capacity for high-density computing platforms. This collaboration focuses on scaling production of systems featuring NVIDIA Blackwell Ultra GPU architectures to meet growing regional AI cloud demand.
Why it matters
By combining Compal’s manufacturing reach with Verda’s cloud operations, the companies aim to accelerate the deployment of high-performance AI infrastructure. This move addresses the need for faster hardware scaling to support localized, high-compute AI workloads.
The SGX30-2 platform utilizes NVIDIA HGX B300 systems, packing 8 NVIDIA Blackwell Ultra GPUs per unit. This configuration targets high-density AI cloud deployments, though total throughput benchmarks remain unverified.
The players
Verda
A Helsinki-based AI cloud operator that has secured over $450 million in total funding and employs 250 people from 40 different countries.
Compal
A Taipei-based electronics manufacturer established in 1984 that builds complex computing platforms for global data center deployments.
The details
Compal leverages its established manufacturing infrastructure to scale production of the SGX30-2, a high-density server platform designed for large-scale AI clusters. The system utilizes NVIDIA HGX B300 boards, which provide the interconnect fabric necessary to link multiple graphics processing units (GPUs) into a unified compute resource. Verda integrates this hardware into its existing AI cloud operations, applying specialized software management to optimize power and compute allocation across these distributed systems.
Timeline
1984: Compal was established.
October 21, 2026: Verda and Compal showcased their expanded collaboration at GTC Berlin.
The Tech Race
This partnership marks a critical push to move the NVIDIA Blackwell platform from announcement to global availability. It follows the industry-wide effort to align high-density manufacturing capacity with the compute requirements of latest-generation AI hardware roadmaps.
The expansion means that regional cloud service providers may soon see reduced lead times for high-density AI infrastructure deployments. Organizations relying on Verda for AI cloud compute will see increased capacity availability as Compal scales its Daxi and U.S. manufacturing lines.
The takeaway
This partnership signals a shift toward regional manufacturing diversification for specialized AI hardware. Analysts should watch for future disclosures regarding the specific output volumes from the new U.S. and Taiwanese assembly facilities to gauge total market availability.
Further reading
For more on the hardware driving massive server farms, visit Data Centers.
Live Poll
Do you believe domestic manufacturing expansion by large technology companies is good for the country?
