High-bandwidth memory (HBM) accounts for 63% of AI chip component costs, up from 52% in Q1 2024. Epoch AI's breakdown of component cost shifts across major chip designers.

Advanced packaging constrained AI chip production in late 2024, followed by HBM memory bottlenecks through 2025. Epoch AI's analysis of semiconductor manufacturing capacity consumed by leading AI chip designers.

Amazon, Google, Meta, Microsoft, and Oracle collectively hold an estimated 71% of the world's cumulative AI compute as of Q4 2025, up from 63% in Q1 2024.

We estimate Google is the largest single owner of AI compute, holding about one quarter of global cumulative capacity as of Q4 2025, primarily from its own custom TPU chips.

We announce our new AI Chip Owners explorer, showing which companies own the world’s leading AI chips.

We estimate that the four largest AI chip designers consumed around 90% of global CoWoS capacity and HBM supply in 2025, while accounting for only ~12% of advanced logic die production.

Total AI data center power capacity reached approximately 30 GW in the last quarter of 2025—comparable to peak power usage in New York State, and outstripping many developed countries.

Total available computing capacity from AI chips across all major designers has grown by approximately 3.3x per year since 2022, enabling larger-scale model development and consumer adoption.

We estimate that NVIDIA’s B200 costs around $6,400 to produce, with High Bandwidth Memory accounting for half of that.

Hyperscalers are planning rapid data center buildouts, with several targeting less than 2 years between the start of construction and reaching 1 gigawatt of operational power.