Nvidia Corp. President and Chief Executive Officer Jen-Hsun Huang announces the EGX Edge supercomputing platform during the company’s event at Mobile World Congress Americas in Los Angeles, California on October 21, 2019.
Patrick T Fallon | Bloomberg | Getty Images
Nvidia’s The most advanced graphics cards sell for more than $40,000 ebayas demand increases for chips needed to train and deploy artificial intelligence software.
Prices for Nvidia’s H100 processors have been noted by 3D gaming pioneers and former Meta Consulting Technology Chief John Carmack on Twitter. At least eight H100s were listed on eBay on Friday, ranging in price from $39,995 to just under $46,000. Some retailers have listed it for around $36,000 in the past.
Announced last year, the H100 is Nvidia’s newest AI flagship chip, succeeding the A100, a roughly $10,000 chip billed as a “workhorse” for AI applications.
Developers use the H100 to create Large Language Models (LLMs), which are at the heart of AI applications like OpenAI’s ChatGPT. These systems are expensive to operate and require powerful computers that process terabytes of data for days or weeks. They also rely on massive computing power for the AI model to generate text, images, or predictions.
Training AI models, especially large ones like GPT, requires hundreds of high-end Nvidia GPUs to work together.
Microsoft Spent hundreds of millions of dollars on tens of thousands of Nvidia A100 chips to develop ChatGPT. Nvidia controls most of the AI chip market.
Nvidia also offers a supercomputer with eight GPUs working together called DGX. Earlier this year, the company announced new services that allow businesses to rent access to DGX computers for $37,000 per month. At this price point, the system would use Nvidia’s older A100 processors.
Nvidia says the H100 is the first chip optimized for the specific AI architecture that underpins many of the recent advances in AI called Transformers. Industry experts say the more powerful chips will be necessary to build even larger and more data-hungry models than those currently available.
REGARD: Google’s chip manufacturing targets Nvidia’s generative AI