Nvidia announced three GPU graphics cards, while AMD also introduced CPUs capable of running AI locally on computers for individual users.
At the event before the CES 2024 Consumer Electronics Show in Las Vegas on January 8, Nvidia introduced three GPU models including RTX 4060 Super, RTX 4070 Ti Super and RTX 4080 Super with prices ranging from 599 to 999 USD. The new GPU series is integrated with a “tensor core” used to run locally generated AI applications, without needing to send information to cloud data.
According to Nvidia representatives, the new GPU series mainly serves users who play games or do graphic design, but can still handle AI applications. For example, RTX 4080 Super can create AI videos 150% faster than the previous generation. All three models are also capable of processing large language models up to five times faster.
They are optimized for each specific software, for example use to create images in Adobe Photoshop’s Firefly creator or remove backgrounds during video calls. Nvidia also created tools that allow game developers to integrate generative AI into games, such as building NPCs (characters in computer-controlled games) that dialogue with players in a more natural way.
Justin Walker, Nvidia senior director of product management, said the new GPU will comply with export controls, so can be sold in China. All three are alternative solutions for researchers and businesses in this country when they cannot access Nvidia’s most powerful server GPU products.
Meanwhile, AMD introduced a trio of new Ryzen 8000G series processors. The most powerful is Ryzen 7 8700G with 8 cores and Radeon 780M GPU, priced at 329 USD; Next is Ryzen 5 8600G with 6 cores, Radeon 760M GPU priced at 229 USD; and the Ryzen 5 8500G also has 6 cores, priced at $179.
According to AMD, the two chips 8700G and 8600G support a separate neural processing unit (NPU) called Ryzen AI. This is considered the only desktop CPU with an NPU, supporting acceleration of AI-related tasks. According to ArsTechnica, this is necessary, especially when Microsoft is adding Windows AI and machine learning features that rely on local processing instead of cloud processing.
In addition to Nvidia and AMD, another chip company, Intel, also introduced the 14th generation “Raptor Lake Refresh” Core product series including 18 processors for computers, with capacities ranging from 35 to 65 W. As the name suggests, this is The refreshed version of the Core Raptor Lake chip series launched last October but with upgraded clock speeds and more energy optimization. However, Intel does not mention AI support features in this series of chips.
CES 2024 will take place from January 9 to 12 in Las Vegas. According to CNBC, this year’s event will feature many technology trends related to AI, especially the integration of this technology into consumer products for end users.
