Nvidia recently announced that it is cooperating with several Taiwanese ODM (Original Design Manufacturing) manufacturers Foxconn, Inventec, Quanta, and Wistron on a partner program concerning high-speed cloud computing, according to a press release by Nvidia on June 26. Under the program, Nvidia will provide ODM partners an early access to its HGX reference architecture, alongside with its GPU (Graphic Processing Unit) technologies and design guidelines, to help them speed up time-to-market. Other than Nvidia's AI supercomputers, the HGX reference architecture is designed for the establishment of hyper-scale data centers and has been used in Microsoft's Project Olympus Initiative and Facebook's Big Basin Systems.