AI hardware · Layer

Chip foundries and fabrication

The factories that actually print the transistors. Cutting-edge AI chips are made on the most advanced process nodes, and the leading-edge foundry business is extraordinarily concentrated.

Key facts

TSMC manufactures the overwhelming majority of leading-edge AI accelerators, including for NVIDIA, AMD, Apple, and many others.
Samsung Foundry and Intel Foundry are the main alternatives pursuing leading-edge nodes, but trail in volume at the cutting edge.
A single advanced fab costs tens of billions of dollars and takes years to build and ramp.
Most leading-edge capacity is in Taiwan and South Korea, which is the central geographic fact of AI hardware.

Where it bottlenecks

Leading-edge fabrication is concentrated in a handful of fabs, mostly in Taiwan, making it both a capacity and a geopolitical chokepoint.

Who dominates it

TSMCSamsungIntel

Companies in this layer

Intel

United States

Integrated device maker pursuing a leading-edge foundry business and AI accelerators.

TSMC

Taiwan

The leading-edge foundry that manufactures the large majority of advanced AI accelerators and owns CoWoS advanced packaging.

All companies across the chain

Explore other layers

GPUs and AI accelerators High-bandwidth memory (HBM)Advanced packaging Lithography and fab equipment Networking and interconnect AI servers and systems Data centers Cooling and thermal management Supply constraints and bottlenecks

Back to the AI hardware map

Let's get to work

Want to build with AI for real?

Beyond the explainer, we design, secure, build and run production AI. Tell us what you have in mind.

Start a project See the offer