Verda and Compal deepen AI server partnership ahead of GTC Berlin 2026
Helsinki-based AI cloud provider Verda and Taiwanese manufacturer Compal Electronics are expanding their collaboration to support Verda's growing AI infrastructure footprint. The companies will present the partnership at GTC Berlin 2026, centered on Compal's SGX30-2 server based on NVIDIA HGX B300 systems.
- Verda and Compal are expanding their collaboration on AI infrastructure and will present it at GTC Berlin 2026.
- The showcase centers on Compal's SGX30-2 server, based on NVIDIA HGX B300 systems with eight NVIDIA Blackwell Ultra GPUs.
- Compal is adding capacity in Daxi, Taiwan and the United States to serve US deployments locally and support Verda's growth in EMEA and Asia.
- Verda says it has raised over $450M and employs around 250 people; the release gives no pricing or performance figures.
- The partnership is on show at the Verda booth, 3059.
Manufacturing reach meets cloud operations
According to the announcement, the collaboration combines Verda's AI cloud expertise with Compal's global manufacturing capabilities and AI server engineering. The companies say demand for AI computing continues to accelerate across Europe and global markets, and that deploying infrastructure quickly and efficiently across regions is becoming increasingly critical for Verda as it adds capacity.
Compal says it continues to invest in its manufacturing footprint, including additional capacity in Daxi, Taiwan and in the United States. The aim is to support customers closer to where AI infrastructure is deployed.
The companies describe a split of roles. Verda contributes expertise in GPU cloud infrastructure, deployment and operations, turning computing capacity into scalable AI cloud services. Compal provides the engineering, manufacturing scale and production footprint needed to turn the latest accelerated computing technologies into deployment-ready infrastructure.
SGX30-2 with NVIDIA Blackwell Ultra GPUs
The hardware at the center of the showcase is Compal's SGX30-2, which is based on NVIDIA HGX B300 systems with eight NVIDIA Blackwell Ultra GPUs. The companies say the platform is designed for high-density AI workloads and gives Verda's expanding infrastructure a next-generation computing foundation.
The release does not give pricing, delivery schedules or performance figures for the server or for the planned deployments.
Compal and Verda say the arrangement creates a more agile path from NVIDIA accelerated computing technology to production AI environments. They claim it will help Verda expand capacity faster while maintaining reliability and scalability.
Regional supply for the US, EMEA and Asia
The announcement ties the partnership to specific regions. U.S. deployments can be supported through local manufacturing in the United States, while Compal's expanded Daxi capacity can support Verda's growth across EMEA and Asia. According to the companies, this should improve deployment flexibility, supply resilience and speed as Verda scales globally.
Jorge Santos, Chief Operating Officer at Verda, said: "Compal gives us the manufacturing reach to bring capacity online closer to our customers, and the engineering depth to do it without trading away quality or speed as we scale our AI infrastructure globally."
Alan Chang, Vice President of the Infrastructure Solutions Business Group (ISBG) at Compal, said the company's investment in global manufacturing is designed to support customers as AI infrastructure moves "from rapid innovation to deployment at scale."
Who the companies are and what happens in Berlin
Verda describes itself as a technology company building cloud infrastructure for AI. It is headquartered in Helsinki, operates across Europe, the US and Asia, employs around 250 people from over 40 nationalities and says it has raised over $450M in total funding. Its investors include Emergence Capital, MUFG Innovation Partners and Supermicro, among others.
Compal was established in 1984 and is headquartered in Taipei. It lists design and production facilities in the United States, Taiwan, China, Vietnam, Mexico, Brazil and Poland, and has expanded into areas such as cloud servers, automotive electronics and smart medical and healthcare.
At GTC Berlin 2026, the companies plan to talk with customers and technology partners about how advanced computing, regional manufacturing and cloud operations can work together to speed up AI deployment. Visitors can find the Verda × Compal collaboration at the Verda booth, 3059.
What it means for customers and the market
For organizations looking for GPU cloud capacity, the announcement signals that Verda is lining up regional manufacturing support for its expansion, with production in the US for American deployments and in Taiwan for EMEA and Asia. Whether that translates into faster availability is a claim by the companies and is not backed by figures in the release. The partnership also shows how server makers and AI cloud providers are coordinating on supply as demand for NVIDIA-based infrastructure grows.
