GMAsia
    🇵🇭Philippines·AI News·10 Oct 2026·via The Manila Times

    New logic of modern data center

    Modern data center architectures are evolving to prioritize efficiency and compute density over physical expansion, driven by increasing demands from AI, advanced analytics, and automation. A comparison cited by AMD's general manager for Asia-Pacific, Alexey Navolokin, suggests that solutions like 5th Gen AMD EPYC processors could reduce server requirements by up to 86%, lower energy consumption by 69%, and decrease total cost of ownership by 41% over three years compared to legacy systems.

    Nexa's Summary

    The core shift in data center design, as described, is away from simply adding more physical servers and space, and towards maximizing the processing capacity within a smaller footprint. This approach aims to consolidate workloads, enabling more applications to run on fewer machines. Such a change directly addresses the growing complexity and scale of AI and advanced analytics tasks, which demand significant computational power.

    This emphasis on efficiency and density has direct implications for operational costs and environmental sustainability. Reduced server counts translate to lower energy consumption and less physical space, which in turn can lead to substantial savings in electricity and real estate. For markets like the Philippines, where operational efficiency is highlighted as a key competitive factor, these cost reductions are presented as particularly relevant.

    Beyond cost, the evolution in data center architecture also prioritizes scalability and adaptability. The ability to quickly adjust infrastructure to varying demands, especially for real-time AI processing, is becoming critical. This 'frictionless scaling' allows companies to deploy new services faster and respond more dynamically to market needs, suggesting that the strategic value of computing infrastructure is increasing beyond mere technical support.

    Share this article

    Original reporting by The Manila TimesWe don't republish, read the full story â†’

    Related reading

    6 stories