NVIDIA GPU: chips but not eliminated: Nvidia can get up to $40 billion worth of business from Elon Musk; this is a malfunction
Elon Musk’s Xai and Nvidia are working together
According to Motley Fool, Musk’s AI startup Xai is building a next-generation supercomputer called Colossus to train its large language model (LLM) Grok, which aims to compete with models like Openai’s Chatgpt. To achieve this, according to the report, the company relies heavily on NVIDIA’s industry-leading bargaining chips.
From 100,000 to 200,000 chips, then more
According to Motley Fool, during the initial development phase, Xai used 100,000 NVIDIA GPUs to develop the Giant System. According to the report, the figure immediately increased by 200,000 chips as the company increased its training capabilities.
Also Read: 15 Years Later, Billionaire Mark Cuban leaves Shark Tank; Here’s what he intends to be next, which will buzz fans
Then, a few months ago, Musk and his team of developers decided to double the security of the GPU, and he said at the time that the next training cluster would be five times larger than the current infrastructure, according to Motley Fool.
This means that the Colossus 2 will require 1 million chips (five times the 200,000 chips mentioned above), and Musk estimates the total cost of the project will be between $25 billion and $30 billion.
New estimate: up to $40 billion
But a new report by Beth Kindig, CEO and chief technical analyst of the I/O fund, shows that the Colossus 2 could cost nearly $40 billion. According to the report, the estimate is based on NVIDIA chips, with an average selling price of between $30,000 and $40,000. Toronto becomes a global blockchain hotspot as Canada stabilizes crypto-rule the United States political deadlock
FAQ
Why does XAI need so many NVIDIA GPUs?
Training large AI models like Grok requires huge computing power provided by NVIDIA’s GPU.
How many chips has XAI used so far?
Xai started with 100,000 Nvidia GPUs and quickly expanded to 200,000 chips according to the Motley Fool.
