Many large market companies are developing artificial ielligence models; Now it seems that Elon Musk, the CEO of X, has been able to steal the competition from others by launching the Colossus processing cluster, the most powerful artificial ielligence training system in the world.
In a post on X, Musk stated that his team at xAI, after 122 days of coinuous effort and using 100,000 Nvidia H100 processors, finally successfully activated the Colossus training system. He also added that the processing power of this cluster will be doubled in the future with the addition of 50,000 H200 processors. Musk had previously announced at the beginning of August that training had begun with this supercluster, which is located in Memphis, USA and was known by the same name.
On the other hand, Nvidia, the manufacturer of the processors of this system, also congratulated him and the xAI team by quoting the post of Mask in X. Colossus iroduced a way to obtain exceptional achievemes in energy efficiency.
In addition, “Kathy Wood”, the CEO of “ARK Invest” company, in a post, considered the launch of Colossus as an effective eve in the world of artificial ielligence and poied to the possibility of new achievemes in the near future.
Colossus, the most powerful artificial ielligence training system in the world
While the strongest OpenAI model has been trained with 80,000 GPUs, Colossus is known as the strongest AI training system due to the number of GPUs it has. The H200 processors prese in this system have 141 GB of HBM3E memory and 4.8 TB of bandwidth.
Despite his high position in the field of artificial ielligence, Musk supports the adoption of laws to preve the possible dangers of this technology. He recely posted that California House Bill 1047 should be passed; A bill that has raised concerns among tech activists about possible limits on artificial ielligence innovation in the United States.
