Source: Content from "
China Business Times
", thank you.
According to industry sources, graphics chip giant NVIDIA will launch the next-generation Ada Lovelace architecture graphics processor (GPU) and use TSMC's 5nm process for mass production. In addition to being first used in the Nintendo OLED version of the Switch game console, the Ada Lovelace architecture GPU will be installed in the next-generation RTX 40 series graphics cards next year. NVIDIA's previous generation Ampere architecture GPU uses Samsung's 8nm foundry for mass production, and the new generation of products will return to TSMC for wafer production.
In the third quarter of last year, Nvidia released its Ampere architecture GPU and RTX 30 series graphics cards for the consumer market. However, due to the full-scale purchase of Bitcoin and other cryptocurrency miners, the prices soared two to three times and it was hard to find a card. Although mainland China recently began to sweep the cryptocurrency mining farms in the country, causing a large number of second-hand graphics cards to be released and sold at a reduced price, due to the strong demand for the stay-at-home economy, Nvidia GPUs and RTX 30 series graphics cards are still in short supply.
According to recent news from insiders, Nvidia will launch a new generation of Ada Lovelace architecture GPUs for the consumer market, and will first use them in Nintendo's upcoming new OLED version of the Switch game console, and will launch RTX 40 series graphics cards next year. The Ada Lovelace architecture GPU will use TSMC's 5nm process, which means that Nvidia's consumer GPU wafer foundry orders will return to TSMC.
In addition, Nvidia originally planned to launch the Hooper architecture GPU after the Ampere architecture, which will adopt chiplet design and multi-chip module (MCM) advanced packaging process to greatly improve artificial intelligence (AI) computing performance. It is reported that Nvidia's Hooper architecture GPU is mainly tailored for the data center market, and its performance can be three times higher than that of the Ampere architecture data center A100 GPU.
Legal persons said that Samsung Foundry's advanced process advancement and capacity expansion were lower than expected, causing Qualcomm and Nvidia to be unable to effectively increase shipments this year to respond to strong market demand, so Qualcomm and Nvidia are clearly returning to TSMC for wafer production next year. Among them, Qualcomm's 5G mobile phone chips have begun to use TSMC's 6nm wafer production in the second half of the year, and will expand the use of TSMC's 5nm and 4nm mass production next year. Nvidia's consumer Ampere architecture GPU uses Samsung Foundry's 8nm production, and the new generation of Ada Lovelace architecture GPU will return to TSMC's 5nm wafer production next year.
AMD's next-generation RDNA 3 architecture GPU will also adopt chiplet design and MCM advanced packaging. Industry sources indicate that different chiplets will be produced using TSMC's 6nm and 5nm processes. Intel's Xe architecture GPU will also be divided into different compute tiles, some of which will be produced using TSMC's 6nm and 5nm processes. For TSMC, it is expected to win all GPU wafer foundry orders next year and become the biggest winner.
*Disclaimer: This article is originally written by the author. The content of the article is the author's personal opinion. Semiconductor Industry Observer reprints it only to convey a different point of view. It does not mean that Semiconductor Industry Observer agrees or supports this point of view. If you have any objections, please contact Semiconductor Industry Observer.
Today is the 2747th content shared by "Semiconductor Industry Observer" for you, welcome to follow.
Semiconductor Industry Observation
"
The first vertical media in semiconductor industry
"
Real-time professional original depth
Scan the QR code
, reply to the keywords below, and read more
Wafers｜ICs｜Equipment
｜Automotive Chips｜Storage｜MLCC｜NVIDIA｜Analog Chips
Reply
Submit your article
and read "How to become a member of "Semiconductor Industry Observer""
Reply
Search
and you can easily find other articles that interest you!
京公网安备 11010802033920号