In the first quarter of 2025, the global data center industry experienced an extraordinary 53% year-over-year surge in capital expenditures (Capex), marking the sixth consecutive quarter of double-digit growth. This unprecedented expansion, as reported by Dell’Oro Group, is primarily driven by the relentless demand for artificial intelligence (AI) infrastructure, especially from hyperscale cloud providers who are racing to deploy next-generation hardware.
Hyperscale cloud operators, including the four largest U.S. providers, have significantly escalated their Capex to support AI workloads. A key factor propelling this investment is the widespread adoption of NVIDIA鈥檚 Blackwell GPU architecture and custom accelerators, which have become essential for training and deploying advanced AI models. Baron Fung, Senior Research Director at Dell’Oro, observed that AI-related spending now dominates hyperscale budgets, with operators prioritizing capacity expansion despite facing supply chain constraints.
According to Dell’Oro, the global data center market is projected to achieve 30% growth in 2025, reaching a total value of $257 billion. High-performance accelerated servers, which constitute over a third of total data center spending, are expected to continue being the fastest-growing segment. This growth is primarily fueled by the dominance of AI servers, particularly those powered by NVIDIA鈥檚 Blackwell platform. Although adoption is mainly concentrated among hyperscale providers, white-label manufacturers have secured over 60% of the server market in the first quarter of 2025, driven by cost efficiency and scalability.
In terms of supplier performance, Dell emerged as the leader in the OEM server market with $6.3 billion in revenue (a 6% year-over-year growth), followed by IEIT Systems. HPE also experienced gains, supported by dual-track growth in both traditional and AI servers.
Despite the presence of tariffs and macroeconomic uncertainties, hyperscale operators have maintained aggressive Capex plans, thanks to their diversified supply chains. However, disparities exist between different market segments. While hyperscale providers continue to expand their capacity, enterprise clients have tempered their spending due to economic caution. Additionally, emerging GPU-as-a-Service (GPUaaS) providers are positioned for rapid growth, tapping into enterprises鈥櫬爊eed for scalable AI resources without requiring heavy upfront investments.
The data center landscape is undergoing significant transformation due to the proliferation of NVIDIA鈥檚 NVL72 platform and other advanced architectures. Analysts highlight that AI-driven upgrades will sustain long-term growth, with Blackwell-class hardware expected to dominate deployments through 2026. Simultaneously, edge computing and sustainability initiatives, such as green data centers powered by renewable energy, are gaining momentum, reflecting the industry’s broader efforts to balance performance with environmental responsibility.
In essence, the surge in Capex during the first quarter of 2025 underscores the transformative impact of AI on data center infrastructure. As both hyperscale providers and enterprises prioritize AI readiness, the industry is set for sustained investment, technological innovation, and market expansion. With global data center spending projected to exceed $1.5 trillion annually by 2032, the race to build the backbone of the AI era is well underway.
