As far as we learned today, xAI has given an energy approval of 150MW to the giant data center on which spy planes are flying. given.
Elon Musk announced that 100 thousand liquid-cooled Nvidia H100 (which has been brought together with the support of xAI, X, Nvidia and some other companies in recent months)Very powerful GPU system focused on artificial intelligence training) announced that he started to work. The latest statement from Elon Musk on the subject is as follows: had happened: “This weekend the xAI team brought online our 100k H100 training cluster, which we call “Colossus”. This system was completed from start to finish in 122 days and Colossus is currently the most powerful artificial intelligence education system in the world.” Two different developments have now brought this system to the agenda. First, xAI received official permission to use a full 150 MW of energy, which will enable all 100,000 GPUs to operate simultaneously. The second development was quite surprising and interesting. As reported, xAI’s rivals in the field are flying spy planes over the ‘Colossus’ data center. It is said that current photos and videos of the center were taken from these planes, but it is not yet known who exactly is behind the process.
YOU MAY BE INTERESTED IN
According to the latest information, this large system will collapse in the next two months. It will start hosting exactly 200 thousand Nvidia H100s. According to a claim, OpenAI CEO Sam Altman recently met with some Microsoft executives and told them that he was concerned that xAI would soon have access to more computing power than OpenAI. Musk previously said, ““The infrastructure will provide a significant advantage in training the world’s most powerful artificial intelligence by every measure by December this year.” The center, which he explained, has strengthened the company’s hand in the training of large language models at the center of LLM, that is, generative artificial intelligence.
Musk had previously said the following on this subject: “The reason we decided to build 100,000 H100s and the next big system in-house is because our competitiveness depends on being faster than other AI companies. This is the only way to bridge the gap. If our fate depends on being the fastest by far, we should be at the wheel rather than a backseat passenger.”