Nvidia taps Samsung for custom chip push

ByEnigma 808

NVIDIA has added Samsung Foundry to its NVLink Fusion ecosystem, signaling a major step in its AI hardware strategy. Announced at the OCP Global Summit, this partnership positions Samsung to provide full design-to-manufacturing support for custom CPUs and XPUs that can integrate with NVIDIA’s NVLink interconnect.

Samsung won’t be building NVIDIA’s own chips but will instead enable third parties to create custom silicon compatible with NVIDIA’s infrastructure. This move supports growing demand from AI-focused firms needing high-speed, scalable compute solutions.

The partnership follows strengthened ties between NVIDIA and Samsung, including HBM3E supply deals. With Samsung’s advanced nodes like SF2 (2nm) drawing interest, it’s clear the company is ready to serve Big Tech’s custom silicon needs.

Author

  • Enigma XO avatar (80x80)

    Besides writing and being a content creator, Enigma 808 sometimes likes to watch paint dry on walls, listening to ASMR while timing snails racing. Such is life.