Hon Hai Technology Group (Foxconn) reaffirmed on October 1 that its next-generation AI servers will begin shipping in the fourth quarter of 2024, with production models on display at the upcoming Hon Hai Technology Day event on October 8 and 9.
Notably, Foxconn will produce 80% of the components and subsystems—excluding CPUs and GPUs—for the AI servers it ships. Servers based on Nvidia's GB200 chip are scheduled to enter mass production in November and begin shipping in December.
As a leading electronics manufacturing services provider, Foxconn occupies a crucial position in the global electronics manufacturing industry, handling over 800 new product introduction (NPI) projects annually. The company's recent developments in the AI server industry further underscore its significance and competitive strength.
Unlike the products shown earlier at Computex 2024, Foxconn emphasizes that the servers to be exhibited at its Technology Day event are actual production versions entering mass production in November, complete with real coolant in the liquid cooling systems.
The company highlights that the GB200 servers demand vastly more complex designs compared to previous server cabinets. For system makers, this translates to much more component-level work rather than simple assembly, a shift that aligns perfectly with Foxconn's strengths.
Beyond supplying 80% of the components for its systems, Foxconn is also reported to be a supplier for the NVLink Switch System, indicating that server cabinets will make a substantial contribution to the company's revenue.
Foxconn confirms that production will proceed as scheduled, with mass production commencing in November and shipments beginning in December. The initial production run in November is expected to focus primarily on NVL72 systems, which will start shipping to clients in early December.
Article translated by Kevin Wang
