Japanese telecom company SoftBank is investing 150 billion yen ($960 million) by 2025 to give its computing facilities the number-crunching power needed to develop world-class generative artificial intelligence. SoftBank aims to develop a Japanese-language-specific generative AI with world-class performance and hopes that the investment will help it achieve this goal. The company plans to purchase graphics processing units (GPUs) from U.S. chip designer Nvidia, which will not only be used for SoftBank’s development of generative AI but will also be loaned to other companies.
The investment in computing infrastructure is believed to be the largest of any Japanese company. SoftBank is currently developing a large language model (LLM), which serves as the basis for generative AI. The company expects to complete development of a model with 390 billion parameters in fiscal 2024, and will start developing a high-performance model with 1 trillion parameters as early as next year. This 1 trillion-parameter model is a de facto benchmark for world-class performance, and is expected to help SoftBank achieve its goal of developing world-class generative AI.
Generative AI is becoming increasingly important in various industries, and the Japanese generative AI market is expected to grow to around $13 billion by 2030, according to the latest forecast by Statista. SoftBank’s move to create its own AI has been well-received by investors, with the company’s share price hitting a record high in late March. Although U.S. tech companies currently lead in terms of both performance and investment, SoftBank’s investment in computing infrastructure is likely to put it among the highest in Japan and help it catch up with its competitors.
Table of Contents
Japanese telecom company SoftBank has announced its plan to invest 150 billion yen ($960 million) by 2025 to enhance its computing facilities for developing world-class generative artificial intelligence models. This investment is believed to be the largest of any Japanese company in computing infrastructure.
Last year, SoftBank invested 20 billion yen in building its computing infrastructure. The additional investment of 150 billion yen will be spent this year and next. The company is confident that the investment will give it the necessary number crunching power to develop a Japanese-language-specific generative AI with world-class performance.
SoftBank’s computing power is expected to be among the highest in Japan when the investment is completed. The company is currently developing a large language model (LLM), which will serve as the basis for its generative AI. The telecom company expects to complete development of a model with 390 billion parameters in fiscal 2024 and will start developing a high-performance model with 1 trillion parameters as early as next year.
SoftBank will purchase graphics processing units (GPUs) from U.S. chip designer Nvidia. The GPUs will not only be used for the company’s own development of generative AI but will also be loaned to other companies. The company hopes that the acquisition of Nvidia chips will help it develop top models.
The stock market is responding positively to SoftBank’s move to create its own AI. Investors are betting that generative AI will be good for business. According to the latest forecast by Statista, the Japanese generative AI market is expected to grow to around $13 billion by 2030, which is 17 times larger than in 2023.
While U.S. tech companies such as Open AI are currently ahead in terms of both performance and investment, Japanese companies, including NTT and NEC, have entered the market. However, the number of parameters in their models runs from several billion to several tens of billions. SoftBank is determined to develop world-class generative AI and hopes to achieve this through its strategic investment in computing infrastructure.
SoftBank is investing 150 billion yen ($960 million) to develop world-class generative artificial intelligence (AI) with a Japanese-language-specific focus. The company aims to boost its computing facilities’ number-crunching power, which is expected to be among the highest in Japan when the investment is completed by 2025.
The company is currently developing a large language model (LLM) that serves as the basis for generative AI. The focus is on developing a Japanese-language-specific generative AI with world-class performance. SoftBank will also loan out the graphics processing units (GPUs) it will purchase from Nvidia to other companies.
SoftBank expects to complete the development of a 390 billion-parameter model by fiscal 2024. The company will start developing a high-performance model with 1 trillion parameters as early as next year. This model is a de facto benchmark for world-class performance, with the number of parameters in GPT-4, developed by OpenAI of the U.S., which performed well on a mock bar exam, reaching the trillion-parameter range.
At present, U.S. tech companies such as OpenAI are ahead in terms of both performance and investment. Japanese companies, including NTT and NEC, have entered the market, but the number of parameters in their models runs from several billion to several tens of billions. SoftBank’s large language model ambitions aim to benchmark against global standards and develop a Japanese-language-specific generative AI with world-class performance.
Generative AI is starting to be integrated into services and products in various industries. The Japanese generative AI market is expected to grow to around $13 billion by 2030, which is 17 times larger than in 2023, according to Statista, a German research specialist. The stock market is responding positively to SoftBank’s move to create its own AI. Investors are betting generative AI will be good for business.
The generative AI market in Japan is expected to grow significantly in the coming years, with Statista projecting it to reach around $13 billion by 2030. This represents a 17-fold increase from the market size in 2023. The growth is driven by the increasing integration of generative AI into services and products across various industries. SoftBank’s investment of $960 million to boost computing power for generative AI is expected to further fuel the growth of the market, as the company aims to develop a Japanese-language-specific generative AI with world-class performance.
SoftBank’s move to create its own AI has had a positive impact on the stock market, with the company’s share price hitting 2,064 yen in late March 2024, the highest since its listing in 2018. Investors are betting that generative AI will be good for business, and SoftBank’s investment in computing infrastructure is believed to be the largest of any Japanese company. While U.S. tech companies such as Open AI are currently ahead in terms of both performance and investment, SoftBank’s investment is expected to position the company as a major player in the generative AI market.
The race to develop generative AI has become a highly competitive field among countries around the world. At present, U.S. tech companies such as OpenAI are leading the way in terms of both performance and investment. OpenAI is known for developing large language models with a high number of parameters, such as GPT-3 with 175 billion parameters, which performed well on various language tasks.
China is also making significant investments in AI, with the government setting a goal to become the world leader in AI by 2030. In 2020, China’s total AI investment reached $15.2 billion, according to a report by McKinsey & Company. The country’s leading tech companies, such as Alibaba and Baidu, are also investing heavily in AI research and development.
SoftBank’s recent investment in computing power for generative AI is a sign of Japan’s growing interest in this field. While Japanese companies, including NTT and NEC, have entered the market, the number of parameters in their models runs from several billion to several tens of billions, which is still far behind the trillion-parameter range.
However, SoftBank’s investment is expected to give the company the computing power needed to develop world-class generative AI, and the company’s share price has already seen an increase as investors bet on the potential success of the project. The Japanese generative AI market is expected to grow to around $13 billion by 2030, according to a forecast by Statista.
Domestically generated AI is also considered important by countries around the world. In Japan, the government has set a goal to create a “Society 5.0” where AI and other advanced technologies are used to solve social problems and improve people’s lives. The government is also promoting research and development in AI through various initiatives, such as the “AI Japan R&D Network” and the “AI Grand Challenge.”
In conclusion, the development of generative AI has become a highly competitive field among countries around the world. While the U.S. and China are currently leading the way, Japan’s recent investment in computing power for generative AI is a sign of the country’s growing interest in this field. Domestically generated AI is also considered important by countries around the world, including Japan, where the government is promoting research and development in AI through various initiatives.
Recently, a senior executive at Baidu, which is China's most popular search engine, resigned from…
The Biden administration has taken another significant step in its efforts to safeguard the American…
In recent years, the China-Pakistan Economic Corridor (CPEC) has emerged as a transformative project, reshaping…
Chinese car dealers are increasingly turning towards homegrown electric vehicles (EVs), according to a report…