Research firm IDC reports that artificial intelligence (AI) infrastructure spending reached $89.7 billion in the first quarter of 2026, with rack-scale GPU servers based on the Arm (ARM.US) architecture displacing x86 processors to become the mainstream accelerated computing platform. x86 processors are manufactured by Intel Corp (INTC.US) and Advanced Micro Devices Inc (AMD.US).

Spending Outlook

IDC has also revised its full-year 2026 AI infrastructure spending forecast upward to $497 billion, representing a year-over-year increase of nearly 56%. This growth is primarily attributed to increased capital expenditure from hyperscale cloud service providers and emerging non-GPU AI demand. The research firm now anticipates the global AI infrastructure market will surpass $1 trillion in 2029, reaching $1.08 trillion, and climb to $1.21 trillion by 2030. This represents a compound annual growth rate (CAGR) of approximately 30% from 2025.

Key Q1 Insights

In the first quarter of 2026, AI infrastructure expenditure surged 33% year-over-year. However, growth is normalizing from a higher base, with spending remaining essentially flat sequentially. The report indicates the Q1 data confirms that AI infrastructure investment has transitioned from an initial proof-of-concept phase into a multi-year capital investment cycle. The industry's competitive focus is shifting from "how much compute to buy" to "which platform will win." Enterprise technology buyers, cloud service providers, and national governments are making long-term decisions regarding infrastructure location, budgets, and which AI workloads to prioritize.

Server-related spending accounted for $87.6 billion, or 97.6% of total AI infrastructure spending in the quarter. An increasing amount of AI demand is being met by infrastructure that does not rely on GPU acceleration at all. Hyperscale data center operators, while expanding GPU clusters, are deploying solutions like AI orchestration tools, data pipeline task clusters, and pure-CPU inference clusters to manage costs.

In Q1 2026, the market size for non-x86 (Arm) accelerated servers jumped to $53 billion, up from $47.5 billion in Q4 2025 and $29.8 billion in Q3 2025. Conversely, the x86 accelerated server market size declined to $34.6 billion in Q1 2026, down from $42.7 billion in Q4 2025 and $51.9 billion in Q3 2025.

Juan Seminara, research director for IDC's Global Infrastructure Tracker, stated, "The Q1 2026 data clearly shows AI infrastructure investment has entered a new phase. The focus is no longer on the volume of compute resources being purchased, but on which platform will win. We see x86 accelerated server sales dropping from $52 billion to $35 billion in just two quarters, while ARM rack sales have nearly doubled. This is not a demand collapse but an architectural shift, though with new x86 platforms on the horizon, the final outcome is not yet determined. Simultaneously, we are seeing truly AI-driven demand emerge, such as pure-CPU inference clusters, AI orchestration tools, and more AI-aware storage upgrades."

The report notes the shift in server market share began in Q4 2025, reflecting consolidation by large buyers around NVIDIA Corp's (NVDA.US) NVL72/GB200 rack-scale platforms and a shift away from custom x86 rack designs. The ultimate market direction depends on the product iteration pace of various vendors; persistent industry supply chain bottlenecks mean it remains to be seen which architectural path will ultimately prevail.

Storage Demand and Regional Performance

The report adds that pent-up storage procurement demand is being released. Over the past year or two, enterprises prioritized budgets for GPUs and AI servers, repeatedly delaying storage equipment refreshes, making further deferrals increasingly difficult. The report also points out that this backlog of storage refresh needs, combined with new, rigid demand driven by AI business, further highlights the urgency of strategic external storage deployment. However, AI-specific storage currently represents a low 2.4% share of total AI infrastructure investment.

Regional performance varied. The United States maintained its dominant position with Q1 spending of $67.9 billion, accounting for 75.7% of global expenditure. This represents a 30.3% year-over-year increase, though growth continues to slow. Spending in China grew 9.3% year-over-year to $7.8 billion, representing 8.7% of the global total. The Middle East region, primarily Saudi Arabia and the UAE, again recorded the world's strongest year-over-year spending growth in Q1 2026, driven by government-backed sovereign AI initiatives and partnerships with leading hyperscale data center operators, though spending declined sequentially from a record Q4 2025.

Seminara commented, "Despite global economic and geopolitical tensions seemingly weighing on other markets, AI investment is demonstrating remarkable resilience." In Wednesday's pre-market trading, Arm shares fell nearly 3%, AMD shares declined about 2%, and Intel shares dropped 3.4%.