Server Shipments Projected to Increase Slightly in 2024, Focus on AI Servers

The driving force behind server shipments in 2024 continues to be America-based Cloud Service Providers (CSPs), according to TrendForce.
Report on global server shipment for 2024However, challenges such as high inflation and elevated corporate financing costs are hindering capital expenditures, resulting in overall demand not yet reaching pre-pandemic levels.

It is estimated that global server shipments will reach around 13.654 million units in 2024, marking a modest increase of approximately 2.05 percent. Notably, the market is witnessing a growing emphasis on the deployment of AI servers, with their shipment share expected to be around 12.1 percent.

Foxconn is anticipated to experience the highest growth rate, estimated at about 5–7 percent annually. This growth is fueled by significant orders from companies such as Dell, AWS (Amazon Web Services) for their Graviton 3 and 4 platforms, Google for Genoa, and Microsoft for Gen9. In terms of AI server orders, Foxconn has secured notable contracts with Oracle and AWS for ASIC orders.

Inventec is predicted to have the second-highest growth rate, expected to increase by about 0–3 percent annually. Despite a decline in OEM orders, CSP orders, particularly from AWS for Graviton 3 and 4, as well as Google for Milan and Genoa, are providing support. Furthermore, Inventec anticipates strong demand from Chinese customers like ByteDance. The company’s AI server shipments are projected to grow at a double-digit rate, constituting approximately 10–15 percent of its total shipments.

Quanta Computer and Supermicro are forecasted to see flat growth rates in server shipments. Quanta faces uncertainties, including a reduction in general-purpose server orders from Meta in the first half of the year and a shift in demand from Intel to AMD platforms by Google. However, Quanta is expected to perform better in AI server shipments, driven by orders from North American cloud customers like Microsoft and AWS.

Supermicro is concentrating on AI server growth, aiming to potentially double its shipments in this segment. However, the lack of a significant rebound in general-purpose server shipments suggests overall levels are likely to remain stable. Supermicro’s high-end AI servers, including models like HGX, cater mainly to second-tier data centers in Europe and the US, such as CoreWeave and Tesla. The company is actively pursuing AI orders from clients including Apple and Meta.

Overall, ODMs (Original Design Manufacturers) anticipate robust AI server shipments in 2024, driven primarily by orders from North American cloud data centers. The growth rate and market share of AI servers are expected to reach double digits, although growth for Chinese clients may face limitations due to US restrictions. Servers equipped with advanced AI training chips, such as NVIDIA’s H series or AMD’s MI series, are poised to experience significant shipment increases.

Baburajan Kizhakedath

Related News

Latest News

Latest News