Qisda Partners with AMD to Launch AI Server Rack-Level Solution

(Central News Agency, Taipei, June 4) The BenQ Qisda Group announced today a partnership with Advanced Micro Devices (AMD) to launch an AI server rack-level solution based on the AMD platform, and has already secured orders from a US customer.

BenQ Qisda Group General Manager Ke Shu-fen stated that the group will not be absent from building AI infrastructure. Driven by the AI business, performance is expected to improve quarter by quarter this year, with full-year results surpassing last year.

Qisda unveiled the server rack at the Taipei International Computer Show (COMPUTEX) today, aiming to achieve top-tier computing power. Giga Computing has adopted the 6th generation AMD EPYC platform to build a premier AI computing infrastructure. Meanwhile, it integrates computing, networking, and storage resources to provide a one-stop 'full-rack delivery' AI solution. The server rack, after technical integration, utilizes Giga Cooling's 2P DLC new two-phase direct liquid cooling technology and Alpha Networks' 1.6T high-speed switches.

Ke Shu-fen emphasized that although Qisda entered the AI race later, 'better late than never.' The current goal is to target global customers, not limited to the Taiwanese market.

BenQ Qisda Group Smart Solutions Business Group General Manager Tseng Wen-hsing pointed out that in addition to the US customer, European customers also have server racks in the validation phase, and orders from Chinese customers have been secured. He explained that the Qisda Group is currently pursuing the AI market with a low-volume, high-mix strategy, rather than mass production.

Tseng Wen-hsing explained that as AI workloads continue to grow, data center infrastructure is shifting from competition in single-machine performance to integrated competition in rack-level system efficiency and energy management. Qisda combines the group's capabilities in servers, liquid cooling, networking, and system integration to meet the demands of AI data centers for high-density computing, low-latency transmission, and energy efficiency management. (Editor: Chang Chun-mao) 1150604