The popularity of chatgpt and other artificial intelligence -based chatbot (AI) is likely to have an impact on the availability of GPU. The reason is, chatgPT is a predictable requires tens of thousands of Nvidia graphics cards.
According to trendforce projections, Openai as the Creator of ChatGPT will require more than 30,000 GPU NVIDIA to process training data. But gamers don't need to worry because the GPU needed is not from the GeForce RTX series.
This projection is made based on the assumption that OpenAI will use the NVIDIA A100 GPU for the Language Model that is chatting. A100 is a computing accelerator that is far more powerful than a commercial graphics card for gaming.
The price is far more expensive than ordinary graphics cards, reaching USD 10,000 to USD 15,000 (around Rp. 153-229 million), depending on the configuration and form factor.
TrendForce claims that chatgpt requires 20,000 graphics cards to process data in 2020. Given its increasingly shot popularity and increasingly widespread use, trendforce estimates that opening will require more than 30,000 graphics cards.
Because the need for graphics cards is getting bigger, Nvidia is expected to share the benefits of the AI chatbot trend. The company from the United States is expected to reap revenue of USD 300 million from the sale of this graphics card.
But the figure can be bigger because Nvidia also has a more powerful H100 graphics card. This graphics card is claimed to have three times the performance of the A100, and the price is more fantastic starting from USD 30,000, as quoted by the Tom's Guide, Friday (3/3/2023).
The GPU market for the data center is not only filled by Nvidia - Intel and AMD also sell AI accelerator. But Nvidia is considered as an optional company to provide AI solutions.
So, are gamers who have difficulty finding graphics cards because they are bought up by crypto miners will again face the scarcity of GPU? It will depend on several factors.
The graphics card needed by OpenAI is not a class of graphics cards for gaming. But if Nvidia diverts production to pursue the needs of the GPU of the Data Center, the supply of consumer class graphics cards can be limited.
