Nvidia has confirmed that all of its Ampere 2nd generation RTX cards are being manufactured using the Samsung 8N node, as opposed to the TSMC 7nm manufacturing process used for the first Ampere GA100 GPU. Rumors have been flying around for months as Nvidia has been trying to decide whether to use TSMC or Samsung for its next generation cards.

Nvidia's Jen-Hsun Huang reportedly added to the question by saying that the company will use TSMC for the majority of its GPUs but will use Samsung's chip manufacturing facilities for a small portion of its graphics silicon production.

Switching to September 1, a special GeForce event was held where Jen-Hsun Huang announced that the massive Nvidia Ampere RTX GPUs would all be manufactured on Samsung's 8N process. How close that is to TSMC's 7nm node will remain to be seen until we actually get our hands on the cards.

However, Nvidia is calling this the "biggest generational shift in history," with performance per watt nearly double that of the still quite fast Turing GPUs.

With smaller Samsung process nodes and twice the rumored number of CUDA cores in each SM, the new generation of GeForce chips will offer enormous rasterizing performance. [The $499 RTX 3070 is reportedly capable of well over 60fps at 1440p with ray tracing enabled. Additionally, the $699 RTX 3080, which Nvidia calls the flagship of its Ampere series despite the existence of the ultra-priced RTX 3090, is said to exceed 4K at 60fps with all RTX features.