What did AI infrastructure spending total in Q1?

Global AI infrastructure spending reached $89.7 billion in the first quarter of 2026, IDC reported. The figure climbed 33.1 percent year over year, lifting the analyst firm's outlook for the rest of 2026 and beyond.

The market covers servers with GPUs and other XPUs, network switches, and a small slice of storage. Processors represent a small share of total system cost, especially in dense rackscale designs.

Did Arm really pass X86 in AI hosts?

Yes, on revenue. Arm-based AI hosts captured 60.5 percent of the segment. X86 systems from Intel and AMD held 39.5 percent. Arm AI revenue ran above X86 AI revenue of $34.6 billion after Nvidia's "Grace" CG100 processors entered NVL72 platforms alongside Hopper and Blackwell GPUs.

Who else runs Arm in AI clusters?

Google has begun deploying its Axion Arm processors inside TPU clusters. Amazon Web Services has run Graviton Arm chips as AI host CPUs for several years. The hyperscalers are following the same path toward cheaper, custom silicon.

How does the broader server market look?

IDC tracked $122.6 billion in total server sales for Q1 2026. X86 systems reached $63.9 billion, a 52.1 percent share. Arm systems hit $56.4 billion, a 46 percent share. IBM's Power and z machines earned an estimated $2.3 billion, a 1.9 percent share, after author extraction from IDC shipment data.

The Arm share grew fastest. It will pass X86 in calendar 2026 if hyperscaler buying patterns hold.

Where is the spending growing fastest?

The United States dominated AI infrastructure purchases at $67.9 billion, up 30.3 percent. China reached $7.8 billion with 9.3 percent growth. Asia Pacific and Japan hit $5.8 billion, expanding 62 percent, twice the U.S. rate. Western Europe grew 20.4 percent to $5.1 billion. The Middle East and Africa crossed $1 billion on a 3.3-times jump from a small base.

How do the long-term forecasts compare?

IDC projects AI infrastructure spending at $1.21 trillion in 2030. The compound annual growth rate from 2023 through 2030 reaches 49.2 percent, close to independent estimates from Nvidia and AMD. AMD chief executive officer Lisa Su described a steeper path at the recent Advancing AI 2026 event in Silicon Valley.

How wide is the gap between the two views?

Scale AMD's AI accelerator total addressable market by server, switch, and storage multipliers, and the implied 2030 AI system revenue reaches $2.11 trillion. The $900 billion gap to IDC widens from 2027 onward as networking, racks, and storage rise faster than accelerator prices.

Non-AI X86 servers still outsold non-AI Arm servers in Q1 2026, at $29.3 billion versus $3.4 billion, but the gap is narrowing.

What should buyers and investors watch next?

Hyperscaler buying patterns will decide whether the market tracks IDC's $1.21 trillion line or AMD's $2.11 trillion line by decade's end. The first test arrives when IDC releases its 2027 outlook and AMD reports next earnings.