9621 Agnes Crossing, Lake Suzanneview, New Mexico Island 84604-9295.
Nvidia is benefiting from the huge AI hype that is driving everyone to invest in GPU accelerators today. OpenAI ignited the spark last year with the launch of
Nvidia is benefiting from the huge AI hype that is driving everyone to invest in GPU accelerators today.
OpenAI ignited the spark last year with the launch of ChatGPT, the iPhone moment of AI, according to Nvidia CEO Jensen Huang. What followed was an avalanche of AI announcements from nearly every company involved in IT. All of these things require a lot of processing power to run on a GPU due to the massive parallel processing power of this architecture compared to the CPU.
Nvidia cannot meet the demand today and is trying to significantly increase the number of orders from the chip manufacturer TSMC. Prices are currently an average of 40 percent above the recommended retail price and bottlenecks are to be expected by the end of this year.
According to DigiTimes, China is also pulling hard on Nvidia’s sleeve. The AI hype is forcing companies like Baidu to buy chips en masse. Due to the trade war between the US and China, Nvidia is not allowed to sell its A100 and H100 GPUs, but the manufacturer offers modified versions called the A800 and H800.
The Nvidia A100 GPU is the “older” model and is baked in the 7-nanometer process. The latest H100 GPU uses the brand new 4 nanometer process. It goes without saying that companies today rely on Nvidia’s H100 GPU and DGX H100 servers with H100 GPUs.
According to Wccftech, Nvidia is initially prioritizing non-Chinese technology partners. This has increased the standard delivery time in China from three months to six months or longer.
On the other hand, GPUs for gamers don’t fly over the counter like hotcakes these days. The artificial bubble created by cryptominers is still pushing prices up too much today, depressing demand. Nvidia has reportedly shifted some of that manufacturing capacity to more data center GPUs to meet high demand there.
Microsoft previously announced that OpenAI trains its GPT models on an HPC cluster in Azure with tens of thousands of GPUs. It was recently announced that Elon Musk also wants to get back into the AI world. To do this, he founded a new company and bought the necessary hardware: another 10,000 GPUs.
Source: IT Daily
As an experienced journalist and author, Mary has been reporting on the latest news and trends for over 5 years. With a passion for uncovering the stories behind the headlines, Mary has earned a reputation as a trusted voice in the world of journalism. Her writing style is insightful, engaging and thought-provoking, as she takes a deep dive into the most pressing issues of our time.
