Top US cloud service providers continue to lead growth in the AI server market, according to TrendForce
In sum – what to know:
AI momentum – Global AI server shipments are projected to rise 24.3% in 2025, slightly below forecasts due to U.S. export restrictions and geopolitics.
Cloud strategies – AWS, Google, Microsoft, Meta and Oracle are expanding AI infra with varying mixes of Nvidia GPUs and in-house chips.
OEM shifts – Server vendors are revising H2 2025 strategies, while sovereign cloud and regional AI projects are boosting demand in EMEA, Asia.
Global demand for AI servers remains strong in 2025, with ‘tier-one’ North American cloud service providers (CSPs) continuing to lead market growth, according to new analysis from market research outfit TrendForce. Risng demand is also being supported by tier-two data centers and sovereign cloud deployments across Europe and the Middle East, it said.
Despite robust momentum for AI servers at a global level, TrendForce has lowered its annual forecast marginally due to growing geopolitical tensions and ongoing U.S. export controls, chiefly affecting the Chinese market. Due to this, the global AI server shipments are now projected to grow 24.3% year-over-year — still well into double-digit territory.
Among the big U.S.-based hyperscalers, Microsoft continues to prioritize AI infrastructure investments, favoring Nvidia GPU-powered systems. The company’s internally developed ASICs — specifically the next-generation Maia chips — are progressing more slowly, with volume ‘ramp’ expected in 2026, according to TrendForce. Microsoft’s AI push has also slowed its purchases of general-purpose servers, it added.
Meanwhile, Meta is expanding both its AI server capacity and general-purpose server footprint. New data center completions have driven strong AMD-based server orders. Meta’s in-house AI chip program is also growing, with MTIA chip shipments expected to double by 2026, according to the report.
TrendForce highlighted that Google’s server growth has been mainly driven by sovereign cloud projects and fresh data center capacity in Southeast Asia. “Google has begun mass deployment of its AI inference-focused TPU v6e chips, which have become mainstream in the first half of 2025,” it said.
Amazon Web Services (AWS) is focused on scaling its in-house Trainium v2 platform while developing several variants of Trainium v3, scheduled for mass production in 2026. AWS is expected to lead all U.S. CSPs in in-house AI chip shipments this year, doubling its 2024 volumes, the research firm found.
“Oracle, compared to the other four CSPs, is more focused on purchasing AI servers and in-memory database (IMDB) servers. In 2025, Oracle plans to step up AI server infrastructure deployment and integrate its core cloud database services with AI applications. In response to sovereign cloud projects in the US, the company has also seen increased demand for NVIDIA’s GB Rack NVL72 solutions,” said TrendForce.
“Additionally, many server enterprise OEMs are reassessing their market strategies for the second half of 2025 due to recent changes in international tariff policies. Currently, TrendForce estimates that total server shipments—including both general-purpose and AI servers—will see a year-on-year growth of approximately 5%, in line with previous projections,” it added.
