Elon Musk’s artificial intelligence (AI) startup, xAI is expecting to build its own supercomputer by 2025. According to the “The Information” report released in investor presentation on May 26, 2024, the tech billionaire, CEO of Tesla and X (twitter) wishes to launch the project — which would run the next repetition of xAI’s chatbot Grok — as soon as next year.
Elon Musk has stated that in public the xAI will need up to 100,000 dedicated semiconductors to train and run the next version of Ai Grok, and will necessitate a “gigafactory of compute” to house those chips.
According to The Information, the supercomputer will integrate Nvidia’s flagship H100 graphics processing units (GPUs) and will be at least four times larger than the largest current GPU clusters in operation today. The report also mentions that xAI might work in partnership with Oracle on the xAI supercomputer project.
The news follows reports that xAI has protected new backing, fetching it close to a $6 billion funding round, which would worth the company at $18 billion.
Talking with investors Elon Musk said that xAI can compete with the likes of OpenAI, Anthropic and Google by beating into data, technology and business of other companies he owned and leads, including Twitter/X, Tesla and Starlink.
The xAI debuted Grok in November-2023. It is a ChatGPT-like AI assistant which boasts an inspiring 314 billion parameters. To place this into viewpoint, OpenAI’s GPT-3 model had 175 billion parameters. In this March, xAI released open-sourcing Grok, a move that some say could democratize AI technology and lead to innovation in commercial applications.
In an interview with PYMNTS “Richard Gardner” the CEO of the tech consultancy Modules said that “Open source models could lead to more rivalry and, hypothetically, lower costs for businesses looking to influence AI in their operations”.
As the original models become more accessible, competition might change towards the quality of data used to train them and the knowledge in applying those models to solve particular business problems.
Further, Richard Gardner added that open source AI allows for faster innovation and more transparency than exclusive models developed by big tech companies. “Open source supports free speech and decreases the potential for suppression by big tech,” Richard said, noting that it also creates AI more available to smaller companies and startups that might not consume the resources to build their models from scratch.
