Softbank to spend $960m on computing infrastructure to develop an ...
Softbank has announced plans to significantly increase its investment in computing infrastructure, with a budget set to rise from $130 million last year to an impressive $960 million for the 2024-2025 period. This move aims to enhance the company's capabilities in artificial intelligence (AI), according to a report by Nikkei.
During the previous year, SoftBank established a new division which has now been elevated to the status of a subsidiary named SB Intuitions. The primary focus of SB Intuitions is the development of a large language model (LLM) featuring an extensive 390 billion parameters, which is expected to be ready for deployment in the fiscal year 2024. To put this into perspective, OpenAI's GPT-3, a comparable model, only boasts 175 billion parameters.
Advancements in LLMs
The number of parameters in an LLM directly impacts its performance, with a greater number of variables resulting in more accurate inferences. Next year, SB Intuitions plans to launch a language model with an impressive 1 trillion parameters, while OpenAI's upcoming GPT-4 is anticipated to feature a staggering 100 trillion parameters.
Expansion of Datacenters
Additionally, SoftBank is laying the groundwork for the construction of data centers across Japan. These facilities will be instrumental in delivering services that leverage internally developed LLMs to cater to the needs of customers.
Stay tuned for more updates on Softbank's ambitious initiatives in the realm of computing infrastructure and AI development.
