laitimes

Nvidia will invest a lot of money to compete for TSMC's production capacity: for the RTX 40 series graphics card

As we all know, Nvidia's RTX 30 series graphics card GPU is handed over to Samsung to OEM production, but there are reports that the RTX 40 series may not be handed over to Samsung again, but transferred to TSMC to OEM.

Nvidia will invest a lot of money to compete for TSMC's production capacity: for the RTX 40 series graphics card

According to foreign media reports, Nvidia has prepaid about $1.64 billion for TSMC in the third quarter of 2021 and will pay $1.79 billion in the first quarter of 2022, and the entire long-term order advance payment will reach $6.9 billion, much higher than the price they paid last year. In addition, Nvidia is said to use this money not only to purchase capacity from TSMC, but also to obtain capacity from Samsung, but it looks like most of the money will be used for TSMC's 5nm technology.

Nvidia will invest a lot of money to compete for TSMC's production capacity: for the RTX 40 series graphics card

This advance payment is much larger than in the same period of previous years, and the purpose of NVIDIA's initiative is to ensure the supply of wafers on TSMC and obtain certain priority in production.

The Ada Lovelace architecture will correspond to the RTX 40 series of graphics cards, which are rumored to be officially released in the fourth quarter of 2022 and will become the "king of NVIDIA's fiscal year 2022".

According to previously released news, the flagship product with Ada Lovelace architecture GPU will be equipped with 18432 CUDA cores, with GDDR6X memory, and has a 384-bit width, and has a core frequency of 2.2 GHz, which can provide 81 TFLOP computing performance (FP32).

Nvidia will invest a lot of money to compete for TSMC's production capacity: for the RTX 40 series graphics card

In addition, the news about the mobile version of the RTX 3080 Ti has also been released recently, according to the news of foreign media Videocardz, NVIDIA's upcoming RTX 3080 Ti will be the first mobile GPU equipped with 16Gbps GDDR6 video memory. It will also have a higher TGP compared to the existing model.

According to reports, the specific specifications of the notebook RTX 3080 Ti are uncertain, and it is known that it will be equipped with up to 16GB of 16Gbps GDDR6 video memory. For comparison, the current notebook RTX 3080 is equipped with 14Gbps GDDR6 video memory.

In addition, the notebook RTX 3080 Ti will have a maximum power consumption of 175W, which is 10W higher than the RTX 3080.

NVIDIA's upcoming notebook RTX 3070 Ti may also be equipped with 16Gbps video memory, and the power consumption may reach 150W.

Read on