Ilon Musk, the founder of SpaceX and Tesla, intends to build a supercomputer called “Calculation Gigatric” to support its XII AI-Startap. The supercomputer, which will consist of 100,000 NVIDIA chips, should be launched by the fall of 2025. “Musk will personally take responsibility for compliance with the terms of implementation of this ambitious project,” reports The Information.
It is expected that the power of the new supercomputer will “at least four times higher than the capabilities of the largest GPU-claster currently existing”, used, for example, META* to teach its AI models. Musk made such a statement during a recent presentation for investors.
Since the appearance of a powerful generative and Openai ChatGPT II-tool, this technology has become the subject of rigid competition between Microsoft and Google technological giants, as well as META and startups like Anthropic and Stability Ai.
Musk is one of the few investors in the world with sufficient financial resources to compete with Openai, Google or Meta in the field of AI. He was one of the co-founders of Openai in 2015, but left the company in 2018, expressing dissatisfaction with its commercial orientation. In March, the billionaire filed a lawsuit against Openai, accusing it of a departure from the initial non -profit mission – to make AI research accessible to everyone. Openai believes that a mask’s lawsuit and his desire to open software development is nothing more than “envy” after leaving the company.
Last year it became known that the XAI is developing a chatbot called GROK. A feature of this product is the ability to turn to the data of the popular social platform X in real time, which distinguishes it among other language models, such as GPT from Openai and Palm from Google.