According to a report by Dell’Oro Group, the remarkable growth was primarily fueled by the increased deployment of Graphics Processing Units , custom accelerators, High Bandwidth Memory , and storage solutions tailored for AI applications. These components have become integral in processing the complex computations required for AI training and inference tasks.
NVIDIA's Hopper platform played a significant role in this expansion, with the introduction of its Blackwell architecture in late 2024 further bolstering the market. NVIDIA captured nearly half of the total market revenues, solidifying its position as a dominant force in the AI hardware landscape.
AMD's Instinct platform also gained traction among major cloud service providers, reflecting a diversification in the accelerated computing market. Additionally, leading cloud giants such as Google, Amazon, and Microsoft significantly increased their adoption of accelerated servers equipped with custom accelerators optimized for AI workloads.
The general-purpose server market experienced a robust rebound, with heightened demand for memory and storage components. Improved pricing throughout the year contributed to this recovery, indicating a revitalized interest in traditional server infrastructure alongside AI-focused investments.
Ethernet adapters witnessed substantial growth, with Smart Network Interface Cards and Data Processing Units revenues increasing by 77%. This uptick is attributed to their widespread deployment in AI training clusters, enhancing data transfer rates and computational efficiency within data centers.
The Central Processing Unit market showed signs of recovery, with revenues projected to increase by 45% in 2025. This resurgence suggests a renewed emphasis on upgrading general-purpose computing capabilities to complement specialized AI hardware.
Looking ahead, continued investments in AI infrastructure are expected to sustain demand for critical components. Projections indicate strong double-digit revenue growth in 2025 for GPUs, custom accelerators, HBM, high-capacity Solid-State Drives , and Network Interface Cards essential for backend networks.
The top ten hyperscalers accounted for more than half of global data center capital expenditures in 2024, reflecting their commitment to expanding AI capabilities. Tier 2 cloud providers also ramped up their investments, approaching hyperscaler levels due to increased GPU deployments.
While NVIDIA maintained its leadership in the data center component market, accounting for nearly half of global revenues, AMD's growing presence and the development of custom accelerators by major cloud service providers indicate a dynamic and competitive landscape.
Topics
Technology