In last year’s AI boom, Nvidia played a central role with expensive server graphics processing units (GPUs), including the H100. They are extremely necessary to train and deploy artificial intelligence generators like OpenAI’s ChatGPT.
Not stopping there, Nvidia continues to demonstrate its power in the consumer GPU segment by announcing three new graphics cards at CES 2024: the RTX 4060 Super, RTX 4070 Ti Super and RTX 4080 Super, priced from $599 to $999. They have additional “tensor cores” designed to run generative AI applications. In addition, Nvidia also provides graphics cards in laptops from Acer, Dell, Lenovo…
Demand for Nvidia’s enterprise GPUs – which cost tens of thousands of dollars per chip and are often used in systems combining 8 GPUs – has helped the company record strong revenue growth, raising its market capitalization to over $1 trillion.
PC gaming graphics cards have long been a cash cow for Nvidia. However, the company says this year’s cards have been improved and geared towards running AI models without having to send information back to the cloud. For example, the RTX 4080 chip can generate AI video 150% faster than its predecessor. Other software improvements make large language models run 5 times faster.
Justin Walker, Nvidia’s senior director of product management, said the company has sold 100 million RTX GPUs. Nvidia hopes next year’s AI applications will take advantage of the power of new GPUs. Microsoft plans to unveil Windows 12, the new version of Windows, later this year.
Nvidia’s new chips can be used to generate images in Adobe Photoshop’s Firefly tool or remove backgrounds when making video calls. Nvidia is also developing tools for game developers to integrate generative AI into their titles.
According to CNBC, the chips show that while Nvidia remains most tied to server GPUs, it will compete with Intel, AMD and Qualcomm in the local AI segment. All three rivals have announced new chips for use in “AI PCs”. The tech industry is figuring out the best way to apply generative AI – which requires enormous computing power and significant expense to run in the cloud.
A technical solution promoted by Microsoft and Nvidia’s rivals is called “AI PC” or sometimes referred to as “edge compute”. Instead of using extremely powerful supercomputers over the Internet, the device will have more powerful onboard AI chips and can run large language models or creative image applications, of course not without shortcomings.
The new graphics cards will comply with US export restrictions and may be sold to China. This is an alternative for Chinese researchers and companies who cannot buy Nvidia’s most powerful server GPUs.
