Marking a Major Breakthrough for the Korean Giant

NVIDIA has now included Samsung Foundry as a partner for custom CPUs and XPUs, with the Korean giant offering its expertise for design and manufacturing.

Custom CPUs & XPUs Made By Samsung Foundry Could Now Leverage NVIDIA’s NVLink Ecosystem

One of the significant news items from the OCP Global Summit was the inclusion of Samsung Foundry in NVIDIA’s growing list of partners for the firm’s AI ambitions, marking a significant breakthrough not just for the Korean giant but also for Jensen himself. In a blog post about OCP, NVIDIA announced that Samsung Foundry has been integrated into the NVLink Fusion ecosystem, where the entity will design custom silicon and provide manufacturing expertise to its customers. While the scope of the partnership hasn’t been defined yet, it is clear that Samsung Foundry has made its way into the NVIDIA camp.

Intel and Samsung Foundry are joining the NVLink Fusion ecosystem that includes custom silicon designers, CPU and IP partners, so that AI factories can scale up quickly to handle demanding workloads for model training and agentic AI inference.

Samsung Foundry has partnered with NVIDIA to meet growing demand for custom CPUs and custom XPUs, offering design-to-manufacturing experience for custom silicon.

The partnership appears not to be limited to manufacturing alone; Samsung will offer ‘end-to-end’ support for custom silicon, encompassing stages such as silicon design, verification, integration, and tape-out. It is revealed that the Korean foundry will meet the “growing demand” for custom CPUs and XePUs, and this partnership will not directly target NVIDIA’s products; instead, it will enable NVIDIA to utilize custom silicon made by Samsung Foundry.

Samsung has grown closer to NVIDIA in recent times, particularly after the meeting between CEO Jensen Huang and Samsung Chairman Lee Jae-yong, following which the Korean giant secured approval for HBM3E supply. The new NVLink Fusion partnership indicates that if firms like OpenAI and others look towards Samsung Foundry for their manufacturing needs, then those chips will be able to leverage NVIDIA’s NVLink ecosystem. This also opens up the prospect for Intel to use the Korean giant’s chip services for its data-center-focused x86 CPUs, but this is unlikely to happen.

The inclusion of Samsung Foundry into NVIDIA’s NVLink ecosystem indicates the readiness of the Korean giant to meet custom silicon demand from Big Tech, driven by processes like the SF2 (2nm), which have garnered massive attention.

https://wccftech.com/nvidia-officially-taps-samsung-foundry-for-custom-silicon/amp/