Global Server Market Hits $166.3B as AI Demand Surges
The global server market reached an unprecedented revenue level in the second quarter of 2026, as artificial intelligence infrastructure spending continued to accelerate across hyperscalers, cloud providers, enterprises, and government-backed AI projects.
According to IDC’s Worldwide Quarterly Server Tracker (Q2 2026), global server vendor revenue reached $166.3 billion in Q2 2026, representing a 52.0% year-over-year increase and a 35.7% quarter-over-quarter increase. The result established a new quarterly revenue record, surpassing the previous high of $125.3 billion recorded in Q4 2025.
The expansion was driven by two forces operating simultaneously: server shipment volumes recovered, while average selling prices (ASPs) continued to rise. AI-accelerated systems were a major contributor, but pricing pressure also affected conventional server infrastructure as memory costs and component allocation constraints tightened supply.
📈 AI Infrastructure Drives Record Server Revenue #
AI infrastructure remained the dominant growth engine for the server industry during Q2 2026.
Hyperscale internet companies and major cloud service providers continued increasing capital expenditure on GPU-optimized infrastructure, while demand expanded beyond traditional public-cloud deployments. Specialized cloud providers, sovereign AI initiatives, enterprises, and government organizations are increasingly purchasing dedicated computing infrastructure for inference, agentic workloads, and other AI applications.
This broadening buyer base is important because it indicates that server demand is no longer dependent exclusively on a small group of hyperscalers.
Q2 2026 Market Snapshot #
| Metric | Q2 2026 Result | YoY Change |
|---|---|---|
| Total Server Vendor Revenue | $166.3 billion | +52.0% |
| Non-x86 Server Revenue | $74.4 billion | +146.0% |
| x86 Server Revenue | $91.9 billion | +16.1% |
| GPU-Accelerated Server Revenue | $87.4 billion | +28.1% |
| Other Accelerated Server Revenue | $27.5 billion | +237.6% |
| Global Server Shipments | — | +15.4% |
| ODM Direct Revenue | $89.7 billion | +35.2% |
Non-x86 systems accounted for 44.8% of total server revenue, narrowing the gap with x86 platforms. Meanwhile, GPU-accelerated servers generated $87.4 billion, equivalent to 52.6% of the overall market.
The figures illustrate how rapidly AI-oriented infrastructure is reshaping the server product mix.
☁️ Cloud CapEx Shows No Sign of Slowing #
Hyperscaler and cloud-provider capital expenditure remained a critical source of demand during the quarter.
Large cloud operators continue deploying AI infrastructure across different geographic regions and workload types, from large-scale model training to inference and increasingly sophisticated agentic AI workloads.
At the same time, the buyer base is becoming more diversified. IDC highlights growing participation from neocloud providers, sovereign AI programs, and enterprises deploying inference and agentic workloads.
This diversification changes the structure of server demand.
Neocloud providers can generate substantial infrastructure volume as they build specialized AI capacity. Sovereign AI programs can be less directly tied to conventional commercial budget cycles, while enterprise AI deployments create potentially longer-term demand as organizations transition experimental workloads into production environments.
From Training Clusters to Inference Infrastructure #
The market is also moving beyond the initial concentration on large AI training clusters.
Inference workloads can require infrastructure to be distributed closer to users and applications, creating demand across additional data centers and geographic regions. Agentic AI workloads introduce another layer of infrastructure requirements because systems may continuously combine model inference with retrieval, orchestration, tool execution, and application-level processing.
As a result, AI server demand increasingly encompasses more than the largest GPU clusters.
💰 Server Shipments Rise as ASPs Surge #
Revenue growth was significantly faster than unit shipment growth.
Global server shipments increased 15.4% YoY, compared with a 52.0% increase in revenue. The difference indicates that higher average selling prices and changes in system configuration contributed substantially to market expansion.
Both accelerated and conventional servers experienced significant pricing increases.
GPU-Accelerated Server Pricing #
GPU-accelerated server ASPs increased from approximately $118,600 to nearly $170,200, representing a 43.6% increase.
Notably, GPU unit shipments declined 10.8% YoY even as accelerated-server revenue continued to grow.
This combination highlights the effect of increasingly expensive AI server configurations. Higher-value GPU systems can generate substantially more revenue per shipment, allowing the market to expand even when unit volumes do not increase at the same rate.
Non-Accelerated Server Pricing #
Conventional server systems also experienced substantial ASP inflation.
Non-accelerated server shipments increased 16.7% YoY, while average selling prices rose from approximately $9,800 to $13,000, an increase of 33.5%.
In this segment, both unit growth and pricing contributed to the increase in revenue.
🧠 Memory Costs and Component Allocation Push Prices Higher #
IDC identifies rising DRAM and NAND flash memory costs and component allocation mechanisms as important contributors to server pricing pressure.
These constraints are encouraging buyers to commit to components and complete systems further ahead of deployment schedules. Such behavior can create additional inventory buffers as organizations attempt to protect themselves against future price increases and supply shortages.
The trend should not necessarily be interpreted as evidence of an equivalent increase in underlying end-user demand. Part of the near-term purchasing activity reflects procurement strategies designed to reduce supply-chain risk.
The constraints also extend beyond components.
Power and Data Center Capacity Become Bottlenecks #
AI infrastructure requires substantially more power and cooling capacity than many conventional server deployments.
As AI clusters become larger and denser, organizations must secure sufficient electrical capacity, cooling systems, networking infrastructure, and physical data-center space. Consequently, the limiting factor for an AI deployment may increasingly be facility readiness rather than the availability of compute hardware alone.
This creates a new execution challenge for server vendors: converting existing orders and committed demand into deployed systems depends on infrastructure that extends well beyond the server itself.
🌎 Regional Server Market Growth Accelerates #
Growth was broad-based across major geographic markets, although the pace varied significantly by region.
| Region | Vendor Revenue | YoY Growth | Share / Notes |
|---|---|---|---|
| United States | $112.2 billion | +54.9% | 67.4% of global revenue |
| China | $26.4 billion | +43.4% | Growth reaccelerated |
| Asia/Pacific ex. Japan | $10.9 billion | +31.5% | — |
| Western Europe | $9.1 billion | +62.7% | — |
| Japan | — | +11.1% | Modest growth |
| Central & Eastern Europe | $0.7 billion | +98.3% | Smaller market base |
| Canada | — | +202.6% | Fastest growth |
| Middle East & Africa | — | +68.8% | — |
| Latin America | — | +32.8% | — |
The United States remained the largest market by a wide margin, generating $112.2 billion in vendor revenue and accounting for 67.4% of global revenue.
Canada recorded the fastest year-over-year growth at 202.6%, although that increase came from a substantially smaller market base. Central and Eastern Europe also posted rapid growth at 98.3%, followed by the Middle East and Africa at 68.8%.
China generated $26.4 billion in server vendor revenue, with growth accelerating to 43.4% YoY.
🏆 Global Server Vendor Rankings #
Branded OEM vendors continued gaining share relative to the ODM Direct model for the second consecutive quarter.
| Rank | Vendor | Q2 2026 Revenue Share | Revenue | YoY Growth |
|---|---|---|---|---|
| 1 | Dell Technologies | 13.4% | $22.24 billion | +165.4% |
| 2 | Supermicro | 6.1% | $10.18 billion | +97.2% |
| 3 | Lenovo | 5.1% | $8.41 billion | +99.6% |
| 4 | HPE | 3.5% | $5.87 billion | +46.0% |
| 5 | Inspur (IEIT Systems) | 2.4% | — | -8.1% |
Dell Technologies retained the top position with 13.4% of global vendor revenue. Its revenue increased 165.4% YoY to approximately $22.24 billion, supported by strong demand for AI server infrastructure.
Supermicro ranked second with 6.1% market share, while Lenovo ranked third at 5.1%. HPE followed at 3.5%, and Inspur ranked fifth with 2.4%.
ODM Direct Market Share Declines #
The ODM Direct model generated $89.7 billion in Q2 2026 revenue, representing 35.2% YoY growth.
Despite this substantial increase, its share of the global market declined from 60.6% in Q2 2025 to 53.9% in Q2 2026.
The shift suggests that branded OEM vendors are capturing a larger proportion of AI infrastructure project spending, even though ODM Direct remains the largest individual market category.
The divergence is notable: ODM Direct revenue continues to expand rapidly, but branded vendors are growing faster in the portions of the market increasingly associated with AI deployments.
🔭 Server Infrastructure Enters a New Growth Phase #
The Q2 2026 results point to a server market being reshaped by both AI-driven volume growth and structural ASP inflation.
Global shipments are recovering, but the revenue increase is substantially larger because modern AI infrastructure carries significantly higher system values. Meanwhile, memory pricing, component allocation, power availability, cooling requirements, and data-center construction timelines are creating constraints throughout the infrastructure stack.
The buyer landscape is also becoming broader. Hyperscalers remain central to demand, but neocloud providers, sovereign AI projects, enterprises, and government organizations are increasingly contributing to infrastructure spending.
For server vendors, the competitive challenge is therefore shifting from simply securing orders to executing them. With large order backlogs already committed, access to components, power, facilities, cooling, and deployment capacity can determine how quickly booked demand becomes realized revenue.
The record $166.3 billion quarterly market therefore represents more than another AI spending spike. It reflects a broader transformation of enterprise computing infrastructure, in which AI workloads are driving higher system complexity, higher ASPs, and increasingly global demand for compute capacity.
Data source: IDC Worldwide Quarterly Server Tracker, Q2 2026, released September 10, 2026. Vendor revenue figures are reported in USD.