AI Chip Market: AI Semiconductor Spending and NVIDIA Data Center Revenue

AI chips include GPUs and other accelerators used to train and run AI models. This report uses Gartner's public forecast for AI processing semiconductors and NVIDIA's investor releases as the clearest company-level indicator.

Gartner put AI processing semiconductor spending at $138.8 billion in 2024 and forecast $209.2 billion in 2025 and $267.9 billion in 2026 [1]. That is growth of about 50.7% in 2025 and 28.1% in 2026 (FREEE.REPORT calculation).

NVIDIA reported record revenue of $215.9 billion for fiscal 2026, up 65% [2]. Data Center revenue rose 68% to a record $193.7 billion [2], about 89.7% of total revenue (FREEE.REPORT calculation).

In the fourth quarter ended January 25, 2026, NVIDIA's Data Center revenue was a record $62.3 billion, up 22% from the previous quarter and 75% from a year earlier [2]. CNBC reported networking sales within Data Center of $10.98 billion in the quarter, up 263% year over year [3]. GAAP gross margin was 71.1% for the fiscal year [2].

Stanford's AI Index reports that AI hardware costs have declined by 30% annually, while energy efficiency has improved by 40% each year [4]. The inference cost for a system performing at the level of GPT-3.5 fell over 280-fold between November 2022 and October 2024 [4].

Governments are funding chip capacity. China launched a $47.5 billion semiconductor fund in 2024 [4].

Gartner figures are end-user spending on AI processing semiconductors in millions of US dollars, converted to billions, from its September 2025 release; 2025 and 2026 are forecasts. NVIDIA figures are reported results for its fiscal year ended January 25, 2026. NVIDIA's Data Center segment includes compute and networking and is not identical to Gartner's AI semiconductor category.

FREEE.REPORT calculations: growth 2025 = $209,192 million / $138,813 million - 1; growth 2026 = $267,934 million / $209,192 million - 1; Data Center share = $193.7 billion / $215.9 billion.