Tesla is accelerating its AI development and plans to allocate $500 million to build a new Dojo supercomputer based on NVIDIA's AI GPU.
Dojo is the electric car maker's supercomputer designed for AI machine learning and video training using video data provided by its vehicle portfolio.
The supercomputer was designed primarily to achieve the goal of fully autonomous cars, which is a big part of Tesla's success.
The company needs massive amounts of computing for artificial intelligence and standard autonomous driving for self-driving cars.
Last year, despite numerous setbacks, the company launched its first supercomputer called the Dojo ExaPod.
The Dojo ExaPod delivers up to 1.1 exaflops of computing power, optimized for AI tasks.
But Elon Musk does not want to stop there because he has great ambitions.
The billionaire revealed that Tesla plans to spend huge sums of money on the new Dojo supercomputer.
Musk explained that the money could be used to purchase about 10,000 H100 AI graphics processors from Nvidia.
Tesla plans to invest more, and this number may eventually reach billions of dollars.
NVIDIA will not receive the full amount as AMD is also expected to receive a portion of the expected payment amount.
In his response to a user, Musk revealed that Tesla is also considering using an AMD chip, but he didn't mention which chip, but it's entirely possible that this could come from the Instinct MI300 series and possibly the MI300X.
Last year, AMD closed the gap with NVIDIA in terms of performance, industrial focus and even display with the introduction of the Instinct MI300 series, capable of running AI software at speeds that rival competitors.
The AI market continues to evolve dynamically as it approaches 2024, and there are signs that the market could grow rapidly as it continues to evolve.
Tesla privately purchased 15,000 H100 GPUs last year, suggesting the company is accelerating its AI efforts in collaboration with NVIDIA.