AI Infrastructure Market: Servers, Cloud, Chips and Devices

AI infrastructure is the hardware and cloud capacity used to train and run AI models, plus AI-enabled devices. Gartner defines it to include AI-optimized IaaS, AI-optimized servers, AI network fabric, AI processing semiconductors and devices [1].

Gartner forecasts AI infrastructure spending of $981.9 billion in 2025, $1,484.4 billion in 2026 and $1,977.7 billion in 2027 [1]. That is growth of about 51.2% in 2026 and 33.2% in 2027 (FREEE.REPORT calculation). Infrastructure accounts for more than half of all forecast AI spending in 2026 [1].

Gartner's September 2025 forecast gives a breakdown of infrastructure-type markets for 2024 and 2025:

• AI-optimized servers (GPU and non-GPU accelerators): $140.1 billion in 2024, $267.5 billion in 2025 [2]
• AI processing semiconductors: $138.8 billion in 2024, $209.2 billion in 2025 [2]
• GenAI smartphones: $244.7 billion in 2024, $298.2 billion in 2025 [2]
• AI PCs: $51.0 billion in 2024, $90.4 billion in 2025 [2]
• AI-optimized IaaS: $7.4 billion in 2024, $18.3 billion in 2025 [2]

These categories overlap in coverage with later Gartner definitions and should be read as a guide to the mix, not added to the September 2026 totals.

NVIDIA reported record fiscal 2026 Data Center revenue of $193.7 billion, up 68%, and record fourth-quarter Data Center revenue of $62.3 billion, up 75% from a year earlier [3]. NVIDIA said its customers are racing to invest in AI compute [3].

Gartner expects hyperscalers and service providers buying AI-optimized servers to remain the largest single area of AI spending [1]. It describes the buildout of AI data centre capacity as the largest infrastructure project humanity has undertaken, and says demand has stayed strong despite rising memory prices [1]. Gartner also notes that AI investment is spreading beyond US tech giants to Chinese companies and new AI cloud providers [2].

Spending figures are from Gartner press release tables in millions of US dollars, converted to billions. 2026 and 2027 values are forecasts; 2025 values are estimates or forecasts depending on the release date. The September 2025 and September 2026 forecasts use different category definitions.

FREEE.REPORT calculations: growth 2026 = $1,484,397 million / $981,920 million - 1; growth 2027 = $1,977,685 million / $1,484,397 million - 1.

NVIDIA figures are from the company's fiscal 2026 results release. Hyperscaler capital expenditure figures from company filings are not yet included in this report.