TAIPEI (Taiwan News) — Cloud service provider Lambda Labs confirmed Thursday that it received its first delivery of the Pegatron-built Nvidia GB200 NVL72 server rack.

The GB200 NVL72 is powered by Nvidia's Grace Blackwell Superchip, which combines two high-performance fifth-generation Blackwell Tensor GPUs and a Grace CPU. The server rack is designed for large language models and generative AI applications.

Founded in 2012, Lambda provides cloud services and infrastructure solutions tailored for engineers and researchers, according to UDN. The company offers Nvidia-powered GPUs accessible via the cloud, with clients that include major tech companies such as Intel and Microsoft, as well as academic institutions like Stanford and Harvard University.

On Friday, Pegatron announced a nearly US$20 million (NT$655 million) investment in Lambda through its subsidiary, Pegatron Investment Co. Ltd., acquiring a 0.76% stake, per CNA. Industry sources revealed that Wistron and Wiwynn have also made investments in Lambda.

Pegatron's AI server portfolio includes the MGX 4U, based on Nvidia's MGX server system and powered by the fourth-generation H200 Tensor Core GPU.